Created
October 23, 2025 12:35
-
-
Save sithumonline/dda80032f37f0e957d97e102e5e646c5 to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| {"authors":["Alavi, SA","Rahimian, A","Mehran, K","Vahidinasab, V"],"type":"Journal Article","year":"2022","title":"Multilayer event-based distributed control system for DC microgrids with non-uniform delays and directional communication","venue":"IET GENERATION TRANSMISSION & DISTRIBUTION","volume":"16","number":"2","pages":"267-281","doi":"10.1049/gtd2.12284","wos_id":"WOS:000693633600001","abstract":"<p>The secondary control layer of microgrids is often modelled as a multi-agent distributed system, coordinated based on consensus protocols. Convergence time of consensus algorithm significantly affects transient stability of microgrids, due to changes in communication topology, switching of distributed generations (DGs), and uncertainty of intermittent energy sources. To minimise convergence time in consensus protocol, this work proposes a multilayer event-based consensus control framework, which is resilient to communication delays and supports plug-and-play (P&P) addition or removal of DGs in DC microgrids of cellular energy systems. A novel bi-layer optimisation algorithm minimises convergence time by selecting an optimal communication topology graph and then adjusts controllers' parameters. Average consensus is achieved among distributed controllers using an event-based consensus protocol, considering non-uniform delays between agents. A realisation method has also been introduced using the directional beamforming technique for topology assignment algorithm based on modern telecommunication technologies. Provided feasibility case study has been implemented on a real-time hardware-in-the-loop (HIL) experimental testbed, to validate the performance of the proposed framework for key purposes of voltage stabilisation and balanced power-sharing in DC microgrids.</p>","id":"81caa5d22f7f","score":8,"tags":["multi-agent","event-driven","graph","orchestration","DC microgrids"],"keep":true,"rationale":"The paper discusses a multi-agent distributed system for consensus control in microgrids, which aligns with the themes of orchestration and event-driven frameworks."} | |
| {"authors":["Arif, TM","Rahim, MA"],"type":"Journal Article","year":"2025","title":"Agentic AI for Real-Time Adaptive PID Control of a Servo Motor","venue":"ACTUATORS","volume":"14","number":"9","doi":"10.3390/act14090459","wos_id":"WOS:001579354600001","abstract":"<p>This study explores a novel approach of using large language models (LLMs) in the real-time Proportional-Integral-Derivative (PID) control of a physical system, the Quanser QUBE-Servo 2. We investigated whether LLMs, used with an Artificial Intelligence (AI) agent workflow platform, can participate in the live tuning of PID parameters through natural language instructions. Two AI agents were developed: a control agent that monitors the system performance and decides if tuning is necessary, and an Optimizer Agent that updates PID gains using either a guided system prompt or a self-directed free approach within a safe parameter range. The LLM integration was implemented through Python programming and Flask-based communication between the AI agents and the hardware system. Experimental results show that LLM-based tuning approaches can effectively reduce standard error metrics, such as IAE, ISE, MSE, and RMSE. This study presents one of the first implementations of real-time PID tuning powered by LLMs, and it has the potential to become a novel alternative to classical control, as well as machine learning or reinforcement learning-based approaches. The results are promising for using agentic AI in heuristic-based tuning and the control of complex physical systems, marking the shift toward more human-centered, explainable, and adaptive control engineering.</p>","id":"8e8a5d30f512","score":6,"tags":["multi-agent","adaptive control","real-time","PID control","AI agents"],"keep":true,"rationale":"The paper discusses the use of AI agents in real-time control, which aligns with the themes of multi-agent orchestration and adaptive systems, though it focuses more on control rather than orchestration."} | |
| {"authors":["Beaver, LE","Malikopoulos, AA"],"type":"Journal Article","year":"2021","title":"An overview on optimal flocking","venue":"ANNUAL REVIEWS IN CONTROL","volume":"51","pages":"88-99","doi":"10.1016/j.arcontrol.2021.03.004","wos_id":"WOS:000664938900018","abstract":"<p>The decentralized aggregate motion of many individual robots is known as robotic flocking. The study of robotic flocking has received considerable attention in the past twenty years. As we begin to deploy flocking control algorithms on physical multi-agent and swarm systems, there is an increasing necessity for rigorous promises on safety and performance. In this paper, we present an overview the literature focusing on optimization approaches to achieve flocking behavior that provide strong safety guarantees. We separate the literature into cluster and line flocking, and categorize cluster flocking with respect to the system-level objective, which may be realized by a reactive or planning control algorithm. We also categorize the line flocking literature by the energy-saving mechanism that is exploited by the agents. We present several approaches aimed at minimizing the communication and computational requirements in real systems via neighbor filtering and event-driven planning, and conclude with our perspective on the outlook and future research direction of optimal flocking as a field.</p>","id":"820c6fede95b","score":8,"tags":["multi-agent","flocking","safety","event-driven","optimization"],"keep":true,"rationale":"The paper discusses robotic flocking, which is relevant to multi-agent systems and emphasizes safety and performance, aligning well with the dissertation's focus on safe orchestration."} | |
| {"authors":["Bo, XY","Chen, XY","Li, HS","Dong, YC","Qu, ZY","Wang, L","Li, Y"],"type":"Journal Article","year":"2021","title":"Modeling Method for the Coupling Relations of Microgrid Cyber-Physical Systems Driven by Hybrid Spatiotemporal Events","venue":"IEEE ACCESS","volume":"9","pages":"19619-19631","doi":"10.1109/ACCESS.2021.3053402","wos_id":"WOS:000615036600001","abstract":"<p>The essence of the microgrid cyber-physical system (CPS) lies in the cyclical conversion of information flow and energy flow. Most of the existing coupling models are modeled with static networks and interface structures, in which the closed-loop data flow characteristic is not fully considered. It is difficult for these models to accurately describe spatiotemporal deduction processes, such as microgrid CPS attack identification, risk propagation, safety assessment, defense control, and cascading failure. To address this problem, a modeling method for the coupling relations of microgrid CPS driven by hybrid spatiotemporal events is proposed in the present work. First, according to the topological correlation and coupling logic of the microgrid CPS, the cyclical conversion mechanism of information flow and energy flow is analyzed, and a microgrid CPS architecture with multi-agents as the core is constructed. Next, the spatiotemporal evolution characteristic of the CPS is described by hybrid automata, and the task coordination mechanism of the multi-agent CPS terminal is designed. On this basis, a discrete-continuous correlation and terminal structure characteristic representation method of the CPS based on heterogeneous multi-groups are then proposed. Finally, four spatiotemporal events, namely state perception, network communication, intelligent decision-making, and action control, are defined. Considering the constraints of the temporal conversion of information flow and energy flow, a microgrid CPS coupling model is established, the effectiveness of which is verified by simulating false data injection attack (FDIA) scenarios.</p>","id":"59a7a0e4d81d","score":7,"tags":["multi-agent","cyber-physical systems","event-driven","safety","modeling"],"keep":true,"rationale":"The paper discusses multi-agent systems in the context of microgrid CPS and event-driven modeling, which aligns with the themes of adaptive orchestration and safety in multi-agent environments."} | |
| {"authors":["Braun, S","Cheng, CT","Dowey, S","Wollert, J"],"type":"Journal Article","year":"2022","title":"Performance Evaluation of Skill-Based Order Assignment in Production Environments With Multiagent Systems","venue":"IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN INDUSTRIAL ELECTRONICS","volume":"3","number":"1","pages":"23-30","doi":"10.1109/JESTIE.2021.3108524","wos_id":"WOS:001371688200004","abstract":"<p>The fourth industrial revolution introduces disruptive technologies to production environments. One of these technologies is multiagent systems (MASs), where agents virtualize machines. However, the agent's actual performances in production environments can hardly be estimated as most research has been focusing on isolated projects and specific scenarios. We address this gap by implementing a highly connected and configurable reference model with quantifiable key performance indicators for production scheduling and routing in single-piece workflows. Furthermore, we propose an algorithm to optimize the search of extrema in highly connected distributed systems. The benefits, limits, and drawbacks of MASs and their performances are evaluated extensively by event-based simulations against the introduced model, which acts as a benchmark. Even though the performance of the proposed MAS is, on average, slightly lower than the reference system, the increased flexibility allows it to find new solutions and deliver improved factory planning outcomes. Our MAS shows an emerging behavior by using flexible production techniques to correct errors and compensate for bottlenecks. This increased flexibility offers substantial improvement potential. The general model in this article allows the transfer of the results to estimate real systems or other models.</p>","id":"d4036cbfd883","score":8,"tags":["multi-agent systems","production environments","event-driven","performance evaluation"],"keep":true,"rationale":"The paper discusses multi-agent systems in production environments, focusing on performance evaluation and flexibility, which aligns well with the themes of orchestration and adaptive systems."} | |
| {"authors":["Cao, L","Cheng, ZJ","Liu, Y","Li, HY"],"type":"Journal Article","year":"2024","title":"Event-Based Adaptive NN Fixed-Time Cooperative Formation for Multiagent Systems","venue":"IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS","volume":"35","number":"5","pages":"6467-6477","doi":"10.1109/TNNLS.2022.3210269","wos_id":"WOS:000869047400001","abstract":"<p>This article focuses on the fixed-time formation control problem for nonlinear multiagent systems (MASs) with dynamic uncertainties and limited communication resources. Under the framework of the backstepping method, a time-varying formation function is introduced in the controller design. To attain the prescribed transient and steady-state performance of MASs, a fixed-time prescribed performance function (FTPPF) is designed and the further coordinate transformation addressing the zero equilibrium point problem is removed. To achieve better approximating performance, a neural network (NN)-based composite dynamic surface control (CDSC) strategy is proposed, where the CDSC scheme is consisted of prediction errors and serial-parallel estimation models. According to the signals generated by the estimation models, disturbance observers are established to overcome the difficulty from approximating errors and mismatched disturbances. Moreover, an improved dynamic event-triggered mechanism and varying threshold parameters are constructed to reduce the signal transmission frequency. Via the Lyapunov stability theory, all the signals in the closed-loop system are semi-globally uniformly ultimately bounded. Finally, the simulation results verify the effectiveness of the developed CDSC strategy.</p>","id":"cad1dfe47d95","score":8,"tags":["multi-agent systems","event-driven","formation control","neural networks","adaptive control"],"keep":true,"rationale":"The paper discusses adaptive control strategies for multi-agent systems, which is relevant to the orchestration of agents in an event-driven context."} | |
| {"authors":["Cardei, I","Mutlu, C","Cardei, M"],"type":"Journal Article","year":"2024","title":"Space-time graph path planner for unsignalized intersection management with a V2V agent coordination architecture","venue":"THEORETICAL COMPUTER SCIENCE","volume":"1020","doi":"10.1016/j.tcs.2024.114871","wos_id":"WOS:001317956800001","abstract":"<p>Reducing traffic congestion and increasing passenger safety are important objectives for emerging automated transportation systems. Autonomous intersection management systems (AIMS) enable large scale optimization of vehicle trajectories with connected and autonomous vehicles (CAVs). We propose a novel approach for computing the fastest waypoint trajectory in intersections using graph search in a discretized space-time graph that produces collision-free paths with variable vehicle speeds that comply with traffic rules and vehicle dynamical constraints. To assist our planner algorithm in decentralized scenarios, we also propose a multi-agent protocol architecture for vehicle coordination for trajectory planning using a vehicle-to-vehicle (V2V) network. The trajectories generated allow a much higher evacuation rate and congestion threshold, with lower O(N) algorithm runtime compared to the state of the art conflict detection graph platoon path planning method, even for large scenarios with vehicle arrival rate of 1/s and thousands of vehicles.</p>","id":"fb810a7c0ce8","score":9,"tags":["multi-agent","orchestration","graph","safety"],"keep":true,"rationale":"The paper discusses a multi-agent protocol for vehicle coordination in trajectory planning, which aligns closely with the themes of multi-agent orchestration and graph-based infrastructure."} | |
| {"authors":["Chen, XL","Xiang, JY","Lu, SF","Liu, YX","He, MG","Shi, DL"],"type":"Journal Article","year":"2025","title":"Evaluating large language models and agents in healthcare: key challenges in clinical applications","venue":"INTELLIGENT MEDICINE","volume":"5","number":"2","pages":"151-163","doi":"10.1016/j.imed.2025.03.002","wos_id":"WOS:001512558500008","abstract":"<p>Large language models (LLMs) have emerged as transformative tools with significant potential across healthcare and medicine. In clinical settings, they hold promises for tasks ranging from clinical decision support to patient education. Advances in LLM agents further broaden their utility by enabling multimodal processing and multitask handling in complex clinical workflows. However, evaluating the performance of LLMs in medical contexts presents unique challenges due to the high-risk nature of healthcare and the complexity of medical data. This paper provides a comprehensive overview of current evaluation practices for LLMs and LLM agents in medicine. We contributed 3 main aspects: First, we summarized data sources used in evaluations, including existing medical resources and manually designed clinical questions, offering a basis for LLM evaluation in medical settings. Second, we analyzed key medical task scenarios: closed-ended tasks, open-ended tasks, image processing tasks, and real-world multitask scenarios involving LLM agents, thereby offering guidance for further research across different medical applications. Third, we compared evaluation methods and dimensions, covering both automated metrics and human expert assessments, while addressing traditional accuracy measures alongside agent-specific dimensions, such as tool usage and reasoning capabilities. Finally, we identified key challenges and opportunities in this evolving field, emphasizing the need for continued research and interdisciplinary collaboration between healthcare professionals and computer scientists to ensure safe, ethical, and effective deployment of LLMs in clinical practice.</p>","id":"21c9ca9bcc70","score":5,"tags":["LLM agents","healthcare","evaluation","multi-agent systems"],"keep":false,"rationale":"While the paper discusses LLM agents in healthcare, it focuses more on evaluation practices rather than the orchestration or event-driven aspects relevant to the dissertation."} | |
| {"authors":["Cintuglu, M","Ishchenko, D"],"type":"Journal Article","year":"2022","title":"Multiagent-Based Dynamic Voltage Support of Power Converters During Fault Ride-Through","venue":"IEEE JOURNAL OF EMERGING AND SELECTED TOPICS IN INDUSTRIAL ELECTRONICS","volume":"3","number":"3","pages":"549-558","doi":"10.1109/JESTIE.2021.3119998","wos_id":"WOS:001371688800018","abstract":"<p>Reference point of applicability (RPA) is a predefined point in a distribution network that performance requirements shall apply. As RPA is not directly controlled by a dedicated distributed energy resource (DER), it requires an aggregate response of multiple DERs over a complex meshed network. This article presents a multiagent-based distributed dynamic voltage support (DDVS) method for coordinated aggregate response of multiple DER units to support RPA voltage profile during and after fault ride-through. DDVS implements a multiagent-based leader target tracking scheme with velocity matching and flock centering features to avoid uncoordinated current injections of cooperative DERs. The DDVS is an event-based approach based on RPA requirements. Two-dimensional voltage support objective is handled: 1) coordinated positive sequence current injection for voltage magnitude boosting and 2) coordinated negative sequence current injection for unbalance voltage mitigation. The results demonstrate that the proposed DDVS algorithm addresses emerging coordinated DVS requirement of DERs.</p>","id":"0e760e5735d9","score":8,"tags":["multi-agent","event-driven","orchestration","dynamic voltage support","distributed energy resources"],"keep":true,"rationale":"The paper discusses a multiagent-based approach for dynamic voltage support, which aligns with the themes of multi-agent orchestration and event-driven systems."} | |
| {"authors":["Córdova-Esparza, DM"],"type":"Journal Article","year":"2025","title":"AI-Powered Educational Agents: Opportunities, Innovations, and Ethical Challenges","venue":"INFORMATION","volume":"16","number":"6","doi":"10.3390/info16060469","wos_id":"WOS:001515866400001","abstract":"<p>Recent advances in large language models (LLMs) have triggered rapid growth in AI-powered educational agents, yet researchers and practitioners still lack a consolidated view of how these systems are engineered and validated. To address this gap, we conducted a systematic literature review of 82 peer-reviewed and industry studies published from January 2023 to February 2025. Using a four-phase protocol, we extracted and coded them along six groups: technical and pedagogical frameworks, tutoring systems, assessment and feedback, curriculum design, personalization, and ethical considerations. Synthesizing these findings, we propose design principles that link technical choices to instructional goals and outline safeguards for privacy, fairness, and academic integrity. Across all domains, the evidence converges on a key insight: hybrid human-AI workflows, in which teachers curate and moderate LLM output, outperform fully autonomous tutors by combining scalable automation with pedagogical expertise. Limitations in the current literature, including short study horizons, small-sample experiments, and a bias toward positive findings, temper the generalizability of reported gains, highlighting the need for rigorous, long-term evaluations.</p>","id":"6393ad8322d6","score":5,"tags":["AI","educational agents","ethical considerations","LLMs"],"keep":false,"rationale":"While the paper discusses AI-powered agents and their orchestration in educational contexts, it lacks a direct focus on multi-agent orchestration or the specific infrastructure aspects relevant to the dissertation topic."} | |
| {"authors":["Costa, DG","Silva, I","Medeiros, M","Bittencourt, JCN","Andrade, M"],"type":"Journal Article","year":"2024","title":"A method to promote safe cycling powered by large language models and AI agents","venue":"METHODSX","volume":"13","doi":"10.1016/j.mex.2024.102880","wos_id":"WOS:001293385600001","abstract":"<p>This paper presents a novel information generation methodology to support safer cycling patterns in urban environments, leveraging for that Large Language Models (LLMs), AI-based agents, and open geospatial data. By processing multiple files containing previously computed urban risk levels and existing mobility infrastructure, which are generated by exploiting open data sources, our method exploits multi-layer data preprocessing procedures and prompt engineering to create easy-to-use, user-friendly assistive systems that are able to provide useful information concerning cycling safety. Through a well-defined processing pipeline based on Data Ingestion and Preparation, Agents Orchestration, and Decision Execution methodological steps, our method shows how to integrate open-source tools and datasets, ensuring reproducibility and accessibility for urban planners and cyclists. Moreover, an AI agent is also provided, which fully implements our method and acts as a proof-of-concept implementation. This paper demonstrates the effectiveness of our method in enhancing cycling safety and urban mobility planning. center dot A novel method that combines LLMs and AI agents is defined to enhance the processing of multi-domain open geospatial data, potentially promoting cycling safety. center dot It integrates urban risk data and cycling infrastructure for a more comprehensive understanding of cycling resources, which become accessible by textual or audio prompts.</p>","id":"a264ec21a345","score":4,"tags":["multi-agent","AI agents","urban mobility","safety"],"keep":false,"rationale":"While the paper discusses AI agents and safety in urban cycling, it does not focus on multi-agent orchestration or event-driven systems relevant to the dissertation topic."} | |
| {"authors":["Di Gennaro, G","Buonanno, A","Fioretti, G","Verolla, F","Pattipati, KR","Palmieri, FAN"],"type":"Journal Article","year":"2022","title":"Probabilistic Inference and Dynamic Programming: A Unified Approach to Multi-Agent Autonomous Coordination in Complex and Uncertain Environments","venue":"FRONTIERS IN PHYSICS","volume":"10","doi":"10.3389/fphy.2022.944157","wos_id":"WOS:000841743500001","abstract":"<p>We present a unified approach to multi-agent autonomous coordination in complex and uncertain environments, using path planning as a problem context. We start by posing the problem on a probabilistic factor graph, showing how various path planning algorithms can be translated into specific message composition rules. This unified approach provides a very general framework that, in addition to including standard algorithms (such as sum-product, max-product, dynamic programming and mixed Reward/Entropy criteria-based algorithms), expands the design options for smoother or sharper distributions (resulting in a generalized sum/max-product algorithm, a smooth dynamic programming algorithm and a modified versions of the reward/entropy recursions). The main purpose of this contribution is to extend this framework to a multi-agent system, which by its nature defines a totally different context. Indeed, when there are interdependencies among the key elements of a hybrid team (such as goals, changing mission environment, assets and threats/obstacles/constraints), interactive optimization algorithms should provide the tools for producing intelligent courses of action that are congruent with and overcome bounded rationality and cognitive biases inherent in human decision-making. Our work, using path planning as a domain of application, seeks to make progress towards this aim by providing a scientifically rigorous algorithmic framework for proactive agent autonomy.</p>","id":"91b24a229271","score":9,"tags":["multi-agent","autonomous coordination","probabilistic inference","path planning","interactive optimization"],"keep":true,"rationale":"The paper presents a unified approach to multi-agent coordination in uncertain environments, which aligns closely with the themes of adaptive orchestration and agent autonomy in the dissertation."} | |
| {"authors":["Ding, KY","Yu, J","Huang, JJ","Yang, YC","Zhang, Q","Chen, HJ"],"type":"Journal Article","year":"2025","title":"SciToolAgent: a knowledge-graph-driven scientific agent for multitool integration","venue":"NATURE COMPUTATIONAL SCIENCE","volume":"5","number":"10","doi":"10.1038/s43588-025-00849-y","wos_id":"WOS:001553727400001","abstract":"<p>Scientific research increasingly relies on specialized computational tools, yet effectively utilizing these tools requires substantial domain expertise. While large language models show promise in tool automation, they struggle to seamlessly integrate and orchestrate multiple tools for complex scientific workflows. Here we present SciToolAgent, a large language model-powered agent that automates hundreds of scientific tools across biology, chemistry and materials science. At its core, SciToolAgent leverages a scientific tool knowledge graph that enables intelligent tool selection and execution through graph-based retrieval-augmented generation. The agent also incorporates a comprehensive safety-checking module to ensure responsible and ethical tool usage. Extensive evaluations on a curated benchmark demonstrate that SciToolAgent outperforms existing approaches. Case studies in protein engineering, chemical reactivity prediction, chemical synthesis and metal-organic framework screening further demonstrate SciToolAgent's capability to automate complex scientific workflows, making advanced research tools accessible to both experts and nonexperts.</p>","id":"b4efafbb148d","score":9,"tags":["multi-agent","orchestration","graph","safety","workflow"],"keep":true,"rationale":"The paper discusses a multi-agent system that utilizes a knowledge graph for orchestrating scientific tools, aligning closely with the themes of adaptive event-driven orchestration and safety in multi-agent environments."} | |
| {"authors":["Domingues, ARP","Hamerski, JC","Amory, AD"],"type":"Journal Article","year":"2021","title":"A fault recovery protocol for brokers in centralized publish-subscribe systems targeting multiprocessor systems-on-chips","venue":"ANALOG INTEGRATED CIRCUITS AND SIGNAL PROCESSING","volume":"106","number":"1","pages":"139-154","doi":"10.1007/s10470-020-01637-6","wos_id":"WOS:000521777800001","abstract":"<p>The publish-subscribe programming model has been an alternative to the design of data-intensive distributed applications in many domains. Recently, this model has been ported to the domain of Multiprocessor Systems-on-Chips, in which applications must use the underlying Network-on-Chip communication infrastructure effectively due to restrictions on the architecture such as low power consumption and limited memory size. In such a scenario, the publish-subscribe model fulfills some of these requirements while providing high-level access to the network hardware to programmers, thus contributing to software quality. However, the publish-subscribe model relies on a single process dedicated to orchestrating the communication at the application level, the broker. Should a broker process crash, the communication between associated nodes may experience delays, downtime, or even inconsistent data. In extreme cases, communication is definitively ruined. Thus, a recovery strategy for brokers in the publish-subscribe model becomes crucial when the application has safety requirements. In this work, we extend a publish-subscribe protocol to add redundancy to brokers' sensitive data. Besides, we provide a recovery protocol to recover brokers in case of a failure. We also provide analytical models to estimate the communication overhead of our approach. We validate our approach in two distinct MPSoC platforms. The results show that our approach inserts a small memory footprint to the system while providing minimal system downtime during recovery.</p>","id":"9c3b0776ffad","score":5,"tags":["publish-subscribe","fault recovery","broker","multi-agent","event-driven"],"keep":true,"rationale":"The paper discusses a fault recovery protocol in a publish-subscribe system, which is relevant to event-driven architectures and could inform safety and orchestration strategies in multi-agent systems."} | |
| {"authors":["Farahani, MA","Haeri, M"],"type":"Journal Article","year":"2024","title":"Decentralised event triggered receding horizon online charge management of electric vehicles","venue":"INTERNATIONAL JOURNAL OF AUTOMATION AND CONTROL","volume":"18","number":"5","doi":"10.1504/IJAAC.2024.140531","wos_id":"WOS:001296736000003","abstract":"<p>In this work, energy management of home customers with electric vehicles and renewable resources is modelled in the form of multi-agent systems. The agent decisions affect the others and the mean field game theory could provide a good solution for decision-making and control in multi-agent systems with a large number of agents. Due to uncertainties in the number of cars, power consumption, and production, online optimisation process is proposed by using the receding horizon concept of predictive control. The main problem in such processes is the calculations each agent should perform every hour. Hence, an event-based optimisation is employed to reduce the computational load. The main contribution of the present work is to optimise electric vehicles charging level in a decentralised and online manner in order to keep the load profile smoother in certain interval while reducing the computational complexity.</p>","id":"7b9acabe3658","score":8,"tags":["multi-agent systems","event-driven","optimisation","electric vehicles","decentralised control"],"keep":true,"rationale":"The paper discusses multi-agent systems and event-based optimisation, which are relevant to the orchestration and adaptive management of agents in the proposed dissertation."} | |
| {"authors":["Farooq, A","Xiang, ZR","Chang, WJ","Aslam, MS"],"type":"Journal Article","year":"2025","title":"Recent Advancement in Formation Control of Multi-Agent Systems: A Review","venue":"CMC-COMPUTERS MATERIALS & CONTINUA","volume":"83","number":"3","pages":"3623-3674","doi":"10.32604/cmc.2025.063665","wos_id":"WOS:001498231200001","abstract":"<p>Formation control in multi-agent systems has become a critical area of interest due to its wide-ranging applications in robotics, autonomous transportation, and surveillance. While various studies have explored distributed cooperative control, this review focuses on the theoretical foundations and recent developments in formation control strategies. The paper categorizes and analyzes key formation types, including formation maintenance, group or cluster formation, bipartite formations, event-triggered formations, finite-time convergence, and constrained formations. A significant portion of the review addresses formation control under constrained dynamics, presenting both model-based and model-free approaches that consider practical limitations such as actuator bounds, communication delays, and nonholonomic constraints. Additionally, the paper discusses emerging trends, including the integration of event-driven mechanisms and AI-enhanced coordination strategies. Comparative evaluations highlight the trade-offs among various methodologies regarding scalability, robustness, and real-world feasibility. Practical implementations are reviewed across diverse platforms, and the review identifies the current achievements and unresolved challenges in the field. The paper concludes by outlining promising research directions, such as adaptive control for dynamic environments, energy-efficient coordination, and using learning-based control under uncertainty. This review synthesizes the current state of the art and provides a road map for future investigation, making it a valuable reference for researchers and practitioners aiming to advance formation control in multi-agent systems.</p>","id":"10a0cddf93c3","score":9,"tags":["multi-agent systems","formation control","event-driven","adaptive control","review"],"keep":true,"rationale":"The paper provides a comprehensive review of formation control in multi-agent systems, which is highly relevant to the orchestration and adaptive mechanisms in the proposed dissertation."} | |
| {"authors":["Garcia, PHS","Luizelli, MC","Rossi, FD"],"type":"Journal Article","year":"2025","title":"Toward real-time IoT multi-sensor data orchestration on wireless sensor networks","venue":"JOURNAL OF SUPERCOMPUTING","volume":"81","number":"13","doi":"10.1007/s11227-025-07749-y","wos_id":"WOS:001556848500001","abstract":"<p>Real-time internet of things applications, such as healthcare monitoring and industrial automation, require immediate data processing, making traditional asynchronous middleware-like queues unsuitable. Efficiently orchestrating heterogeneous sensors-including continuous, event-driven, and query-driven types-is a key challenge due to varying priority levels and resource constraints. This paper proposes RT-IMO, a real-time multi-sensor orchestration strategy that integrates priority-based, load-balanced, and adaptive scheduling strategies to ensure low latency and efficient resource allocation. The proposed RT-IMO dynamically adjusts scheduling priorities to balance latency, fairness, and resource constraints in real time. Designed for high-volume, distributed IoT environments, RT-IMO is suitable for deployment on edge-cloud platforms that require low-latency decisions and scalable scheduling under high data ingestion rates-making it well-aligned with the goals of real-time and high-performance systems. Experimental results demonstrate that RT-IMO improves responsiveness, fairness, and system efficiency compared to existing approaches. Future research will explore machine learning-based adaptive scheduling and its extension to heterogeneous edge computing environments. The results show that RT-IMO ensures low latency, efficient resource use, and fairness, prioritizing critical data while selectively dropping lower-priority tasks. It adapts dynamically to workload variations, outperforming static approaches in responsiveness and stability under high load.</p>","id":"fb2e24925750","score":8,"tags":["IoT","multi-sensor","orchestration","event-driven","real-time","adaptive scheduling"],"keep":true,"rationale":"The paper discusses real-time orchestration strategies for heterogeneous sensors, which aligns well with the themes of adaptive event-driven systems and resource management in multi-agent environments."} | |
| {"authors":["Gui, XB","Lv, HL","Wang, X","Lv, LT","Xiao, Y","Wang, L"],"type":"Journal Article","year":"2025","title":"Enhancing hepatopathy clinical trial efficiency: a secure, large language model-powered pre-screening pipeline","venue":"BIODATA MINING","volume":"18","number":"1","doi":"10.1186/s13040-025-00458-5","wos_id":"WOS:001508336300001","abstract":"<p>Background Recruitment for cohorts involving complex liver diseases, such as hepatocellular carcinoma and liver cirrhosis, often requires interpreting semantically complex criteria. Traditional manual screening methods are time-consuming and prone to errors. While AI-powered pre-screening offers potential solutions, challenges remain regarding accuracy, efficiency, and data privacy.</p><p>Methods We developed a novel patient pre-screening pipeline that leverages clinical expertise to guide the precise, safe, and efficient application of large language models. The pipeline breaks down complex criteria into a series of composite questions and then employs two strategies to perform semantic question-answering through electronic health records: (1) Pathway A, Anthropomorphized Experts' Chain of Thought strategy; and (2) Pathway B, Preset Stances within an Agent Collaboration strategy, particularly in managing complex clinical reasoning scenarios. The pipeline is evaluated on key metrics including precision, recall, time consumption, and counterfactual inference-at both the question and criterion levels.</p><p>Results Our pipeline achieved a notable balance of high precision (e.g., 0.921, criteria level) and good overall recall (e.g., similar to 0.82, criteria level), alongside high efficiency (0.44s per task). Pathway B excelled in high-precision complex reasoning (while exhibiting a specific recall profile conducive to accuracy), whereas Pathway A was particularly effective for tasks requiring both robust precision and recall (e.g., direct data extraction), often with faster processing times. Both pathways achieved comparable overall precision while offering different strengths in the precision-recall trade-off. The pipeline showed promising precision-focused results in hepatocellular carcinoma (0.878) and cirrhosis trials (0.843).</p><p>Conclusions This data-secure and time-efficient pipeline shows high precision and achieves good recall in hepatopathy trials, providing promising solutions for streamlining clinical trial workflows. Its efficiency, adaptability, and balanced performance profile make it suitable for improving patient recruitment. And its capability to function in resource-constrained environments further enhances its utility in clinical settings.</p>","id":"0a0f92e0ccc4","score":3,"tags":["AI","clinical trials","pre-screening","language models"],"keep":false,"rationale":"The paper focuses on AI-powered pre-screening in clinical trials rather than multi-agent orchestration or event-driven systems, making it less relevant to the dissertation topic."} | |
| {"authors":["Guo, SY","Pan, YN","Li, HY"],"type":"Journal Article","year":"2025","title":"Dynamic event-driven optimal consensus control for state-constrained multiagent zero-sum differential graphical games","venue":"APPLIED MATHEMATICS AND COMPUTATION","volume":"484","doi":"10.1016/j.amc.2024.128979","wos_id":"WOS:001295762300001","abstract":"<p>In this paper, a dynamic event-driven optimal control scheme is proposed for the zero-sum differential graphical games in nonlinear multiagent systems with full-state constraints. Initially, to address the dual demands of optimality and state constraints, a set of system transformation functions are introduced to satisfy the state constraints of the agents. Then, by applying the principle of differential game theory, the distributed optimal control problem affected by external disturbances is formulated as a zero-sum differential graphical game, and the performance index function related to neighbor informations and disturbances is designed for each follower. Afterwards, to enhance the utilization of communication resource, a novel dynamic event- triggered mechanism characterized by a dynamic threshold parameter and an auxiliary dynamic variable is developed, which not only exhibits greater flexibility but also diminishes the frequency of triggers. Furthermore, the approximate optimal control strategies are obtained by employing an event-driven adaptive dynamic programming algorithm. Ultimately, a simulation example is presented to verify the applicability of the proposed control approach.</p>","id":"c0f6f7ee741b","score":9,"tags":["multi-agent","event-driven","graph-based","optimal control","dynamic systems"],"keep":true,"rationale":"The paper discusses dynamic event-driven control in multi-agent systems, which aligns closely with the themes of adaptive orchestration and event-driven mechanisms in the dissertation."} | |
| {"authors":["Guo, SY","Pan, YN","Li, HY"],"type":"Journal Article","year":"2025","title":"Dynamic event-driven optimal consensus control for state-constrained multiagent zero-sum differential graphical games","venue":"APPLIED MATHEMATICS AND COMPUTATION","volume":"484","doi":"10.1016/j.amc.2024.128979","wos_id":"WOS:001295762300001","abstract":"<p>In this paper, a dynamic event-driven optimal control scheme is proposed for the zero-sum differential graphical games in nonlinear multiagent systems with full-state constraints. Initially, to address the dual demands of optimality and state constraints, a set of system transformation functions are introduced to satisfy the state constraints of the agents. Then, by applying the principle of differential game theory, the distributed optimal control problem affected by external disturbances is formulated as a zero-sum differential graphical game, and the performance index function related to neighbor informations and disturbances is designed for each follower. Afterwards, to enhance the utilization of communication resource, a novel dynamic event- triggered mechanism characterized by a dynamic threshold parameter and an auxiliary dynamic variable is developed, which not only exhibits greater flexibility but also diminishes the frequency of triggers. Furthermore, the approximate optimal control strategies are obtained by employing an event-driven adaptive dynamic programming algorithm. Ultimately, a simulation example is presented to verify the applicability of the proposed control approach.</p>","id":"c0f6f7ee741b","score":9,"tags":["multi-agent","event-driven","graph-based","optimal control","dynamic systems"],"keep":true,"rationale":"The paper discusses dynamic event-driven control in multi-agent systems, which aligns closely with the themes of adaptive orchestration and event-driven mechanisms in the dissertation."} | |
| {"authors":["Guo, ZJ","Ren, HR","Li, HY","Huang, TW"],"type":"Journal Article","year":"2025","title":"Event-Based Optimal Containment Control for Constrained Multiagent Systems Using Integral Reinforcement Learning","venue":"IEEE TRANSACTIONS ON CONTROL OF NETWORK SYSTEMS","volume":"12","number":"1","pages":"609-619","doi":"10.1109/TCNS.2024.3510353","wos_id":"WOS:001449683500029","abstract":"<p>An optimal event-driven containment control problem is studied for partially unknown nonlinear multiagent systems with input constraints and state constraints. Its novelty lies in the optimization of the performance index while ensuring constraints handling abilities on states and inputs. First, an improved discounted cost function is constructed, and the state and input constraint information are encoded into the cost function by barrier functions and nonquadratic utility functions, respectively. Then, the approximate distributed optimal containment control policy is derived by an integral reinforcement learning (IRL)-based adaptive critic design, where the IRL technique can overcome the limitation of known drift dynamics in previous results. In critic neural networks learning, the weight tuning law is presented by virtue of the concurrent learning technique, which relaxes the persistence of excitation conditions by storing appropriate historical data. In order to reduce the amount of information transmitted through the controller-to-actuator channel, a containment error-dependent dynamic event-triggered mechanism is defined. Theoretical results indicate that signals in closed-loop systems driven by event-triggered optimal controllers are uniformly ultimately bounded, and Zeno behavior is avoided. Finally, the effectiveness of the developed method is illustrated by a simulation example on multiple single-link robot manipulators.</p>","id":"70ea0ba999d9","score":9,"tags":["multi-agent systems","event-driven","reinforcement learning","control systems","constraints"],"keep":true,"rationale":"The paper addresses event-driven control in multi-agent systems with a focus on constraints, which is highly relevant to the dissertation's theme of adaptive orchestration."} | |
| {"authors":["Hartung, T"],"type":"Journal Article","year":"2025","title":"AI, agentic models and lab automation for scientific discovery - the beginning of scAInce","venue":"FRONTIERS IN ARTIFICIAL INTELLIGENCE","volume":"8","doi":"10.3389/frai.2025.1649155","wos_id":"WOS:001569217400001","abstract":"<p>Until recently, the conversation about generative artificial intelligence in science revolved around the textual prowess of large language models such as GPT-3.5 and the promise that they might one day draft a decent literature review. Since then, progress has been nothing short of breathtaking. We now find ourselves in the era of multimodal, agentic systems that listen, see, speak and act, orchestrating cloud software and physical laboratory hardware with a fluency that would have sounded speculative in early 2023. In this review, I merge the substance of our 2024 white paper for the World Economic Forum Top-10-Technologies Report with the latest advances through mid-2025, charting a course from automated literature synthesis and hypothesis generation to self-driving laboratories, organoid intelligence and climate-scale forecasting. The discussion is grounded in emerging governance regimes-notably the European Union Artificial Intelligence Act and ISO 42001-and is written from the dual vantage-point of a toxicologist who has spent a career championing robust, humane science and of a field chief editor charged with safeguarding scholarly standards in Frontiers in Artificial Intelligence. I argue that research is entering a \"co-pilot to lab-pilot\" transition in which AI no longer merely interprets knowledge but increasingly acts upon it. This shift promises dramatic efficiency gains yet simultaneously amplifies concerns about reproducibility, auditability, safety and equitable access.</p>","id":"6dec03114e4d","score":8,"tags":["multi-agent systems","orchestration","AI governance","lab automation","safety"],"keep":true,"rationale":"The paper discusses agentic systems and orchestration in lab automation, aligning well with the themes of multi-agent orchestration and safety in the dissertation."} | |
| {"authors":["Hiriyanna, S","Zhao, WB"],"type":"Journal Article","year":"2025","title":"Multi-Layered Framework for LLM Hallucination Mitigation in High-Stakes Applications: A Tutorial","venue":"COMPUTERS","volume":"14","number":"8","doi":"10.3390/computers14080332","wos_id":"WOS:001559578100001","abstract":"<p>Large language models (LLMs) now match or exceed human performance on many open-ended language tasks, yet they continue to produce fluent but incorrect statements, which is a failure mode widely referred to as hallucination. In low-stakes settings this may be tolerable; in regulated or safety-critical domains such as financial services, compliance review, and client decision support, it is not. Motivated by these realities, we develop an integrated mitigation framework that layers complementary controls rather than relying on any single technique. The framework combines structured prompt design, retrieval-augmented generation (RAG) with verifiable evidence sources, and targeted fine-tuning aligned with domain truth constraints. Our interest in this problem is practical. Individual mitigation techniques have matured quickly, yet teams deploying LLMs in production routinely report difficulty stitching them together in a coherent, maintainable pipeline. Decisions about when to ground a response in retrieved data, when to escalate uncertainty, how to capture provenance, and how to evaluate fidelity are often made ad hoc. Drawing on experience from financial technology implementations, where even rare hallucinations can carry material cost, regulatory exposure, or loss of customer trust, we aim to provide clearer guidance in the form of an easy-to-follow tutorial. This paper makes four contributions. First, we introduce a three-layer reference architecture that organizes mitigation activities across input governance, evidence-grounded generation, and post-response verification. Second, we describe a lightweight supervisory agent that manages uncertainty signals and triggers escalation (to humans, alternate models, or constrained workflows) when confidence falls below policy thresholds. Third, we analyze common but under-addressed security surfaces relevant to hallucination mitigation, including prompt injection, retrieval poisoning, and policy evasion attacks. Finally, we outline an implementation playbook for production deployment, including evaluation metrics, operational trade-offs, and lessons learned from early financial-services pilots.</p>","id":"ddfe82ee6b59","score":5,"tags":["LLM","hallucination","safety","framework","multi-agent"],"keep":false,"rationale":"While the paper discusses safety in LLMs, it focuses on hallucination mitigation rather than multi-agent orchestration or event-driven systems."} | |
| {"authors":["Hu, QL","Shi, YX","Wang, CL"],"type":"Journal Article","year":"2021","title":"Event-Based Formation Coordinated Control for Multiple Spacecraft Under Communication Constraints","venue":"IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS","volume":"51","number":"5","pages":"3168-3179","doi":"10.1109/TSMC.2019.2919027","wos_id":"WOS:000640749000045","abstract":"<p>This paper addresses the relative position coordinated control problem for spacecraft formation flying under an undirected communication graph, whilst considering mass uncertainties, external disturbances, and limited communication resources. A new event-triggered information transmission mechanism is first presented, where each spacecraft only requires accessing to the states of neighbors intermittently. Subsequently, a novel event-based coordinated control scheme is proposed by combining a smooth adaptive projection rule that confines the parameter estimations to well-defined bounded convex hypercubes. Under the proposed control framework, the information exchange among spacecraft occurs only when the specified event is triggered, thereby significantly reducing the communication load and saving the onboard resources. Furthermore, a positive lower bound on interevent time intervals is guaranteed to exclude Zeno behavior. By virtue of Lyapunov stability analysis and graph theory, it is proved that the relative position tracking errors can converge to small invariant sets around the origin, and that all closed-loop signals are bounded, even in the presence of mass uncertainties and external disturbances. Finally, numerical simulations are given to evaluate the effectiveness and highlight the advantages of the developed control algorithm.</p>","id":"aa8f7be611a7","score":8,"tags":["multi-agent","event-driven","graph theory","control systems"],"keep":true,"rationale":"The paper discusses event-driven control mechanisms for multi-agent systems (spacecraft), which aligns well with the dissertation's focus on adaptive event-driven orchestration."} | |
| {"authors":["Hu, QL","Shi, YX","Wang, CL"],"type":"Journal Article","year":"2021","title":"Event-Based Formation Coordinated Control for Multiple Spacecraft Under Communication Constraints","venue":"IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS","volume":"51","number":"5","pages":"3168-3179","doi":"10.1109/TSMC.2019.2919027","wos_id":"WOS:000640749000045","abstract":"<p>This paper addresses the relative position coordinated control problem for spacecraft formation flying under an undirected communication graph, whilst considering mass uncertainties, external disturbances, and limited communication resources. A new event-triggered information transmission mechanism is first presented, where each spacecraft only requires accessing to the states of neighbors intermittently. Subsequently, a novel event-based coordinated control scheme is proposed by combining a smooth adaptive projection rule that confines the parameter estimations to well-defined bounded convex hypercubes. Under the proposed control framework, the information exchange among spacecraft occurs only when the specified event is triggered, thereby significantly reducing the communication load and saving the onboard resources. Furthermore, a positive lower bound on interevent time intervals is guaranteed to exclude Zeno behavior. By virtue of Lyapunov stability analysis and graph theory, it is proved that the relative position tracking errors can converge to small invariant sets around the origin, and that all closed-loop signals are bounded, even in the presence of mass uncertainties and external disturbances. Finally, numerical simulations are given to evaluate the effectiveness and highlight the advantages of the developed control algorithm.</p>","id":"aa8f7be611a7","score":8,"tags":["multi-agent","event-driven","graph theory","control systems"],"keep":true,"rationale":"The paper discusses event-driven control mechanisms for multiple agents (spacecraft) under communication constraints, which aligns well with the themes of adaptive orchestration and event-driven systems in multi-agent environments."} | |
| {"authors":["Hu, XR","Guo, HT","Lao, KW","Hao, JK","Liu, FR","Ren, ZY"],"type":"Journal Article","year":"2025","title":"MiniRocket-MARL synergy for storm tide resilience: MESS-DV enhanced recovery in coastal distribution networks","venue":"APPLIED ENERGY","volume":"401","doi":"10.1016/j.apenergy.2025.126710","wos_id":"WOS:001572118900001","abstract":"<p>In coastal cities, distribution networks are vulnerable to storm tide-induced damage, which can trigger widespread power outages and economic losses if not promptly addressed. However, research addressing this challenge remains scarce. To resolve post-storm-tide fault identification and recovery issues, this study proposes a unified fault identification-recovery framework. Acknowledging the transient nature and spatial uncertainty of busbar faults during disasters, a miniRocket-based fault locator for distribution networks was developed. Validated in a digital twin system utilizing real-world grid data, it achieved 99.69 % accuracy in identifying 16 threephase short-circuit fault locations. An uncertainty simulation environment was established to incorporate multi-system couplings in coastal cities, including storm tide risk zones, power grids, transportation networks, and urban drainage systems. By integrating the synergistic effects of active and passive drainage on low-lying areas and introducing physical-information security constraints related to water immersion depth, an event-driven multi-agent reinforcement learning (MARL) framework was designed for coordinated dispatch of mobile energy storage system (MESS) and drainage vehicle (DV) in post-disaster grid recovery. Testing demonstrated that scenarios incorporating active drainage reduced total power restoration time by approximately 6 h compared to passive-only approaches within simulation constraints. Across all test scenarios, coupled systems achieved full power restoration within 7 h, with no subsequent outages following restoration.</p>","id":"0eda2ca4171d","score":6,"tags":["multi-agent","event-driven","resilience","coastal networks"],"keep":true,"rationale":"The paper discusses a multi-agent framework for disaster recovery, which aligns with the themes of orchestration and event-driven systems, though its primary focus is on resilience in coastal distribution networks."} | |
| {"authors":["Huang, JX","Masroor, S","Ali, ZA"],"type":"Journal Article","year":"2024","title":"A network control system for solving a speed coordination problem in a networked multi-motor drive","venue":"INTERNATIONAL JOURNAL OF MODELLING IDENTIFICATION AND CONTROL","volume":"45","number":"1","doi":"10.1504/IJMIC.2024.141667","wos_id":"WOS:001325892600003","abstract":"<p>A control system connected to sensor and actuator via communication network plays pivotal role in a today's emerging world. The problem of obtaining a consensus in a group of networks connected agents is one of the major areas of research in the network control systems. In industry, a multi-motor system is very much in demand due to common load driven capacity, and cost saving. Coordinated speed plays vital role to control the in-flight movement of multi-rotor UAV/drones, producing hovering, tilting or other necessary flight control movements. Thus, this study uses a leaderless multi-agent consensus model to achieve coordinated control of network connected motor drives such that all the drives reach identical speed. Moreover, this study also incorporates event-based control, so that the continuous time controller update can be avoided, thus offering energy saving. To ensure stable system design, the Lyapunov stability criteria are used, while the obtained design is simulated in MATLAB. The simulated results endorse the design concept, such that the system attains a consensus on motor speed along with energy saving.</p>","id":"3f0940108ba0","score":8,"tags":["multi-agent","event-driven","control systems","consensus","UAV"],"keep":true,"rationale":"The paper discusses a multi-agent consensus model and event-based control, which are relevant to the orchestration and adaptive event-driven aspects of the dissertation."} | |
| {"authors":["Huang, Y","Meng, ZY","Sun, J"],"type":"Journal Article","year":"2024","title":"Distributed Constrained Optimization for Second-Order Multiagent Systems via Event-Based Communication","venue":"IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS","volume":"54","number":"9","pages":"5317-5326","doi":"10.1109/TSMC.2024.3405453","wos_id":"WOS:001249189500001","abstract":"<p>This article studies the distributed constrained optimization problems for the discrete-time second-order multiagent systems (MASs), in which each agent privately owns local cost function and nonidentical convex set constraints. To solve this problem, a projection-based distributed event-triggered algorithm is developed via the constant step-sizes, which achieves an ergodic convergence rate O(1/k) for the general convex functions. By applying the event-triggered mechanism, the proposed algorithm can avoid unnecessary communication among the agents. Moreover, it is shown that the introduced event-triggered component does not sacrifice the convergence rate. Finally, a simulation example is carried out to demonstrate the theoretical results.</p>","id":"a596a828a639","score":9,"tags":["multi-agent systems","event-driven","optimization","communication"],"keep":true,"rationale":"The paper focuses on event-based communication in multi-agent systems, which is highly relevant to the adaptive event-driven orchestration theme of the dissertation."} | |
| {"authors":["Kojchev, S","Hult, R","Kneissl, M","Fredriksson, J"],"type":"Journal Article","year":"2025","title":"A Computation Decomposition Strategy for Optimization-Based Coordination of Automated Vehicles in Confined Sites","venue":"IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS","volume":"26","number":"1","pages":"558-568","doi":"10.1109/TITS.2024.3492029","wos_id":"WOS:001362261300001","abstract":"<p>Coordinating the motion profiles of fully automated vehicles in confined sites presents a challenge, particularly in avoiding conflicts in MUTually EXclusive (MUTEX) zones like intersections, merge-splits, and narrow roads. Prior methods have utilized optimization-based formulations, where the scheduling (order) of vehicles in MUTEX zones is found using a heuristic, followed by solving a nonlinear program (NLP) to determine the optimal state and control trajectories of each vehicle given the schedule. This paper presents a method that reduces the computational demand in the second step by decomposing the NLP into multiple smaller, parallelly solvable, NLPs. The method uses the confined area's road network geometry and the vehicle positions to identify non-significant MUTEX relations and leverage results from graph theory to use this information to identify independent subproblems. We demonstrate that the solution obtained using this approach is equivalent to that obtained by solving the original NLP. Furthermore, by utilizing the dual variables connected to the MUTEX enforcing constraints, the method is capable of a further, sub-optimality-inducing, subdivision of the problem, enabling a trade-off between optimality and computation. We show how the method can be utilized, both when the plan needs to be initially computed and when there is a need for updating an existing motion plan. Simulation examples demonstrate the computational improvement with respect to the non-decomposed problem.</p>","id":"cb91ad92485a","score":7,"tags":["multi-agent","orchestration","graph","safety"],"keep":true,"rationale":"The paper discusses optimization-based coordination of automated vehicles, which relates to multi-agent orchestration and graph-based methods, making it relevant for the dissertation."} | |
| {"authors":["Li, L","Tan, RJ","Fang, JW","Xue, JR","Lv, C"],"type":"Journal Article","year":"2025","title":"LLM-augmented hierarchical reinforcement learning for human-like decision-making of autonomous driving","venue":"EXPERT SYSTEMS WITH APPLICATIONS","volume":"294","doi":"10.1016/j.eswa.2025.128736","wos_id":"WOS:001530664200001","abstract":"<p>Reinforcement Learning (RL) has shown great promise for autonomous driving decision-making. However, such data-driven methods inherently struggle to be deployed in real-world due to their limited generalization to rare but safety-critical scenarios and low sample efficiency, resulting in high computational costs. To address these challenges, we propose a hierarchical RL framework augmented with a large language model (LLM), to enhance decision-making in complex driving environments through semantic understanding and commonsense knowledge. Inspired by human drivers, the LLM serves as an expert high-level planner that interprets textual descriptions of driving scenarios to generate a long-term goal point, a recommended meta-action, and a corresponding explanation, thereby navigating complex environments effectively. To meet real-time requirements, the high-level LLM module operates at a reduced frequency, balancing reasoning capability and inference latency. At the low level, however, it remains challenging for the RL agent to learn a sequence of continuous short-term actions, acceleration and steering, that can achieve the high-level goal while ensuring safety and efficiency. To bridge this gap, we introduce a Goal Gradient-based Transfer (GGT) mechanism that embeds an explicit gradient toward the LLM-generated goal, facilitating efficient policy learning. Additionally, to align the learned behaviors with human intents, we incorporate a human-in-the-loop reward design process.Specifically, the LLM contributes to reward design by generating structurally diverse functions, which are iteratively optimized using expert preferences over RL-generated trajectory pairs to ensure alignment with human values and safety. Overall, experimental comparisons in the CARLA simulator demonstrate that the proposed framework significantly improves generalization, interpretability, and human alignment in diverse and unseen driving scenarios.</p>","id":"00e6e2965048","score":6,"tags":["multi-agent","LLM","autonomous driving","decision-making","safety"],"keep":true,"rationale":"The paper discusses a hierarchical RL framework with LLM integration for decision-making, which is relevant to multi-agent systems and safety in complex environments, though it focuses on autonomous driving rather than orchestration."} | |
| {"authors":["Li, MY","Guo, ST","Chen, C","Chen, CL","Liao, XF","Guan, XP"],"type":"Journal Article","year":"2024","title":"DecAge: Decentralized Flow Scheduling for Industrial 5G and TSN Integrated Networks","venue":"IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING","volume":"11","number":"1","pages":"543-555","doi":"10.1109/TNSE.2023.3301879","wos_id":"WOS:001139144400101","abstract":"<p>The integration of fifth-generation (5G) and time-sensitive networking (TSN) is a key technology to promote the wireless upgrade of manufacturing industry, and some preliminary enhancements on 5G with TSN-integration have been introduced to empower industrial internet ubiquitous connectivity and real-time forwarding. However, at present, the 5G system is only served as a logical bridge with 5G-specific procedures hidden from TSN networks, and thus the deep integration of 5G and TSN is still an open issue. This article focuses on the co-design problem of flow scheduling in 5G and TSN, where 5G grant-free configured-grants (CGs) and TSN cyclic queuing and forwarding mechanisms are adopted. Firstly, we innovatively propose to achieve seamless 5G+TSN transmission by intelligently selecting CGs in 5G, where the starting time offsets of different CGs will influence the flow injection time in the downstream TSN network to mitigate the queue overflow issue. Then, an age-of-information (age) aware decentralized and deterministic scheduling (DecAge) scheme based on actor-critic reinforcement learning is proposed. DecAge guides time-sensitive flow hosts to autonomously make packet sampling-transmission decisions so that timeliness-based age constraints can be respected via collision-free 5G+TSN transmission. DecAge contains multiple local actors learning hierarchically-structured scheduling policies for flow hosts and a central graph attention network-based critic to estimate these policies. Through simulation, DecAge shows the robustness to environmental non-stationarity by elaborately coordinating distributed learning and yields improved performance over non-hierarchical agents on sample-efficient learning.</p>","id":"1f0f081c6344","score":7,"tags":["multi-agent","orchestration","graph","reinforcement learning"],"keep":true,"rationale":"The paper discusses decentralized flow scheduling using reinforcement learning in a network context, which relates to multi-agent systems and orchestration, though it focuses more on network integration than direct agent orchestration."} | |
| {"authors":["Li, QB","Lin, WZ","Liu, Z","Prorok, A"],"type":"Journal Article","year":"2021","title":"Message-Aware Graph Attention Networks for Large-Scale Multi-Robot Path Planning","venue":"IEEE ROBOTICS AND AUTOMATION LETTERS","volume":"6","number":"3","pages":"5533-5540","doi":"10.1109/LRA.2021.3077863","wos_id":"WOS:000655244300004","abstract":"<p>The domains of transport and logistics are increasingly relying on autonomous mobile robots for the handling and distribution of passengers or resources. At large system scales, finding decentralized path planning and coordination solutions is key to efficient system performance. Recently, Graph Neural Networks (GNNs) have become popular due to their ability to learn communication policies in decentralized multi-agent systems. Yet, vanilla GNNs rely on simplistic message aggregation mechanisms that prevent agents from prioritizing important information. To tackle this challenge, in this letter, we extend our previous work that utilizes GNNs in multi-agent path planning by incorporating a novel mechanism to allow for message-dependent attention. Our Message-Aware Graph Attention neTwork (MAGAT) is based on a key-query-like mechanism that determines the relative importance of features in the messages received from various neighboring robots. We show that MAGAT is able to achieve a performance close to that of a coupled centralized expert algorithm. Further, ablation studies and comparisons to several benchmark models show that our attention mechanism is very effective across different robot densities and performs stably in different constraints in communication bandwidth. Experiments demonstrate that our model is able to generalize well in previously unseen problem instances, and that it achieves a 47% improvement over the benchmark success rate, even in very large-scale instances that are x100 larger than the training instances.</p>","id":"1e54fcd50e1d","score":9,"tags":["multi-agent","graph","path planning","GNN","decentralized systems"],"keep":true,"rationale":"The paper discusses a novel approach using Graph Neural Networks for decentralized multi-agent path planning, which is highly relevant to the orchestration of multi-agent systems."} | |
| {"authors":["Li, XL","Chen, C","Lyu, Y","Xie, K"],"type":"Journal Article","year":"2021","title":"Event-based resilience to DoS attacks on communication for consensus of networked Lagrangian systems","venue":"INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL","volume":"31","number":"6","pages":"1834-1850","doi":"10.1002/rnc.5013","wos_id":"WOS:000629012300004","abstract":"<p>In this article, we consider the resilience analysis of the even-based consensus control of networked Lagrangian systems under denial-of-service (DoS) attacks, which are usually implemented by the attackers to block the communication channels. Comparing with the linear cyber-physical systems, the secure control of the networked Lagrangian systems is more complex and challenging, especially under the conditions of asymmetry communication and discontinuous control. To this end, an event-based controller is first designed for consensus control of networked Lagrangian systems in the absence of DoS attacks. Sufficient conditions are given to stabilize the closed-loop systems. Then the resilience analysis is presented for the event-based controller under DoS attacks. Some conditions associated with the DoS duration and frequency are proposed for the control parameters against the attacks. Then Zeno behaviors are proved to be nonexistent in the proposed control scheme. An algorithm is also given to guide the control design.</p>","id":"862130a5c992","score":8,"tags":["multi-agent","event-driven","resilience","DoS attacks","control systems"],"keep":true,"rationale":"The paper discusses event-based consensus control in multi-agent systems, which is relevant for understanding adaptive event-driven architectures in the context of resilience against attacks."} | |
| {"authors":["Lin, CH","Kuo, CF"],"type":"Journal Article","year":"2025","title":"Roles and potential of Large language models in healthcare: A comprehensive review","venue":"BIOMEDICAL JOURNAL","volume":"48","number":"5","doi":"10.1016/j.bj.2025.100868","wos_id":"WOS:001587447500002","abstract":"<p>Large Language Models (LLMs) are capable of transforming healthcare by demonstrating remarkable capabilities in language understanding and generation. They have matched or surpassed human performance in standardized medical examinations and assisted in diagnostics across specialties like dermatology, radiology, and ophthalmology. LLMs can enhance patient education by providing accurate, readable, and empathetic responses, and they can streamline clinical workflows through efficient information extraction from unstructured data such as clinical notes. Integrating LLM into clinical practice involves user interface design, clinician training, and effective collaboration between Artificial Intelligence (AI) systems and healthcare professionals. Users must possess a solid understanding of generative AI and domain knowledge to assess the generated content critically. Ethical considerations to ensure patient privacy, data security, mitigating biases, and maintaining transparency are critical for responsible deployment. Future directions for LLMs in healthcare include interdisciplinary collaboration, developing new benchmarks that incorporate safety and ethical measures, advancing multimodal LLMs that integrate text and imaging data, creating LLM-based medical agents capable of complex decisionmaking, addressing underrepresented specialties like rare diseases, and integrating LLMs with robotic systems to enhance precision in procedures. Emphasizing patient safety, ethical integrity, and human-centered implementation is essential for maximizing the benefits of LLMs, while mitigating potential risks, thereby helping to ensure that these AI tools enhance rather than replace human expertise and compassion in healthcare.</p>","id":"436b7e91ba0c","score":4,"tags":["Large Language Models","Healthcare","AI Integration","Ethics"],"keep":false,"rationale":"While the paper discusses LLMs and their integration into workflows, it does not focus on multi-agent orchestration or the specific event-driven and graph-based infrastructure relevant to the dissertation topic."} | |
| {"authors":["Liu, D","Li, LW"],"type":"Journal Article","year":"2024","title":"A resilient control algorithm to distributed optimization under DoS attacks","venue":"INTERNATIONAL JOURNAL OF ROBUST AND NONLINEAR CONTROL","volume":"34","number":"15","pages":"10346-10364","doi":"10.1002/rnc.7520","wos_id":"WOS:001261673000001","abstract":"<p>This article focuses on a distributed optimization problem under an undirectd and connected network affected by DoS attacks. First, new event-triggered mechanisms are developed to schedule the communication between agents. Then, an event-based resilient distributed optimization algorithm that removes the requirement of continuous communication in the existing results is designed to achieve consensus optimization. Assisted by the design of regulator equations and the algebraic graph theory, it proves that the proposed algorithm is exponentially convergent. Moreover, Zeno behavior can be excluded by reductio ad absurdum. The simulation results validate that the proposed algorithm is effective.</p>","id":"8be0c91c282e","score":8,"tags":["multi-agent","event-driven","graph theory","distributed optimization","resilience"],"keep":true,"rationale":"The paper discusses a distributed optimization algorithm for multi-agent systems that incorporates event-triggered mechanisms and graph theory, aligning well with the themes of adaptive event-driven orchestration."} | |
| {"authors":["Liu, S","Sun, JY","Zhang, HG","Zhai, MN"],"type":"Journal Article","year":"2022","title":"Coordination for Lure Multiagent Systems: Fully Distributed Event-Driven Approach With Single-Event Monitoring Condition","venue":"IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS","volume":"69","number":"12","pages":"4919-4923","doi":"10.1109/TCSII.2022.3191092","wos_id":"WOS:000922028300059","abstract":"<p>This brief studies the event-driven consensus problem for Lur'e multi-agent systems (MASs) under directed leader-follower networks. The proposed algorithm does not contain global information, and each agent only communicates with neighboring agents; thus, it can work in a fully distributed manner. Under directed graphs, existing fully distributed consensus algorithms rely on multiple event detection mechanisms. The proposed algorithm contains only one event-driven condition for each agent and thus has lower device cost and complexity. Finally, the simulation verifies the effectiveness of theoretical results.</p>","id":"6b4373997a43","score":8,"tags":["multi-agent systems","event-driven","distributed algorithms","graph theory"],"keep":true,"rationale":"The paper discusses a fully distributed event-driven approach for multi-agent systems, which aligns well with the dissertation's focus on adaptive event-driven orchestration."} | |
| {"authors":["Liu, S","Zhang, HG","Sun, JY"],"type":"Journal Article","year":"2025","title":"Timer-based distributed coordination for achieving asymptotic consensus in directed communication networks","venue":"INFORMATION SCIENCES","volume":"703","doi":"10.1016/j.ins.2025.121886","wos_id":"WOS:001422932500001","abstract":"<p>This paper addresses the consensus problem of linear multi-agent systems (MASs) in directed graphs using a timer-based event-triggered control algorithm. The proposed distributed algorithm allows each agent to update its control law and event monitoring condition using only relative state information from neighboring agents at discrete event instants. This algorithm minimizes reliance on global information and significantly reduces communication overhead, thereby enhancing both efficiency and scalability. A common challenge in event-based control algorithms is the potential for Zeno behavior, where an infinite number of events could occur within a finite time, making the system impractical. While conventional algorithms avoid Zeno behavior by ensuring nonzero time intervals between events, they often fail to address the issue of excessively short event intervals. Our algorithm overcomes this limitation by establishing a strictly positive lower bound for the interval between events for each agent, thereby not only avoiding Zeno behavior but also ensuring practical applicability and robustness of the control strategy. Through simulation studies, we validate the efficacy of our algorithm in achieving asymptotic consensus in linear MASs over directed graphs.</p>","id":"403ce2a5200a","score":9,"tags":["multi-agent systems","event-driven","graph theory","consensus","distributed coordination"],"keep":true,"rationale":"The paper focuses on distributed coordination in multi-agent systems using event-triggered control in directed graphs, which is highly relevant to the dissertation's focus on adaptive event-driven orchestration."} | |
| {"authors":["Liu, XH","Wang, XY","Fan, BJ","Xiao, GX","Wen, SP","Chen, BD","Wang, P"],"type":"Journal Article","year":"2025","title":"Multiagent Primal-Dual DDPG-Based Reactive Power Optimization of Active Distribution Networks via Graph Reinforcement Learning","venue":"IEEE INTERNET OF THINGS JOURNAL","volume":"12","number":"15","pages":"32058-32071","doi":"10.1109/JIOT.2025.3575445","wos_id":"WOS:001547293500006","abstract":"<p>The large-scale integration of distributed energy resources into active distribution networks (ADNs) may significantly intensify voltage fluctuations and increase network losses. Traditional model-based reactive power optimization (RPO) approaches depend on existence of accurate system models. On the other hand, conventional reinforcement learning (RL) methods largely ignore the spatial characteristics of the ADNs during training, allowing agents to have an inadequate perception of the system state. To address these challenges, this article proposes a multiagent deep RL approach that integrates graph learning with RL for the learning of RPO strategies in ADNs. Specifically, the active distribution network is divided into multiple regions, with each region being controlled by an agent. The agents collaborate to achieve the global RPO goal. The perception capability of the agents is enhanced by adopting graph attention networks during the feature extraction phase. In the training phase, a primal-dual method is employed to manage constraints effectively. During the execution phase, each agent controls the photovoltaic inverters, electric springs, and capacitor banks based on the strategies developed in the training phase. The performance of the proposed approach is validated by a series of experiments on the IEEE-33 system, along with comparisons versus some existing data-driven deep RL methods.</p>","id":"93e0de7fd078","score":8,"tags":["multi-agent","graph-based","reinforcement learning","orchestration","power optimization"],"keep":true,"rationale":"The paper discusses a multi-agent approach using graph learning for reactive power optimization, which aligns well with the themes of multi-agent orchestration and graph-based infrastructure."} | |
| {"authors":["Lu, SD","Yao, YT","Luo, B","Yu, ZF","Li, DL","Shi, WS"],"type":"Journal Article","year":"2022","title":"EdgeWare: toward extensible and flexible middleware for connected vehicle services","venue":"CCF TRANSACTIONS ON HIGH PERFORMANCE COMPUTING","volume":"4","number":"3","pages":"339-356","doi":"10.1007/s42514-022-00100-4","wos_id":"WOS:000792535300001","abstract":"<p>The dramatic development of Edge Computing technologies is strongly stimulating the adoption of machine learning models on connected and autonomous vehicles (CAVs) so that they can provide a variety of intelligent onboard services. When multiple services running on the resource-constrained CAVs, how limited resources can dynamically support the desired services is of the utmost importance for both automakers and domain researchers. In this context, efficiently and dynamically managing vehicle services becomes critical for autonomous driving. While previous research focused on service scheduling, computation offloading, and virtual machine migration, we propose EdgeWare, an extensible and flexible middleware to manage the execution of vehicle services, which is open-source to the community with four key features: i) on-demand model switch, i.e., easily switch and upgrade machine learning models, ii) function consolidation and deduplication to eliminate duplicate copies of repeating functions and maximize the reusability of vehicle services, iii) build event-driven applications to reduce workload, and iv) dynamic workflow customization which enables customizing workflow to extend the functionality. Our experiment results show that EdgeWare accelerates the execution of services about 2.6 x faster compared to the silo approach and save CPU and memory utilization up to around 50% and 17% respectively, and it allows domain researchers to dynamically add new services on CAVs or easily switch to the upgraded applications for the life cycle management of vehicle services.</p>","id":"2b9225e17aa5","score":7,"tags":["middleware","event-driven","autonomous vehicles","service orchestration"],"keep":true,"rationale":"The paper discusses a middleware for managing services in connected vehicles, which aligns with the themes of orchestration and event-driven applications relevant to multi-agent systems."} | |
| {"authors":["Luo, XY","Fu, YL","Li, XL"],"type":"Journal Article","year":"2023","title":"Resilient Synchronization of Networked Lagrangian Systems Over Event-Based Communication With Asynchronous DoS Attacks","venue":"IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING","volume":"10","number":"6","pages":"3198-3208","doi":"10.1109/TNSE.2023.3252666","wos_id":"WOS:001089300800008","abstract":"<p>A wide range of actual systems can be modeled as Euler-Lagrange dynamics, its inherently complex nonlinearities present additional difficulties in the design of control algorithms. In this article, the distributed resilient synchronization control problem of networked Euler-Lagrange (E-L) systems under event-based communication with distributed denial-of-service (DoS) attacks is considered. A novel distributed dynamic event-triggered scheme is proposed to schedule the communication source under asynchronous DoS attacks on different channels. Then, a self-triggered scheme is designed to reduce the updating number of the control signals. Under the proposed adaptive control scheme, the asymptotic synchronization of the closed-loop system is guaranteed under DoS attacks. Neither the control strategy nor the dual-terminal event-triggered scheme needs the eigenvalue information of the Laplace matrix. Also, no Zeno behavior occurs under the proposed event-based control and communication framework. Finally, case studies are provided to show the effectiveness of the proposed method.</p>","id":"0d615eb112ec","score":8,"tags":["multi-agent","event-driven","synchronization","control systems"],"keep":true,"rationale":"The paper discusses event-based communication and synchronization in networked systems, which is relevant to adaptive event-driven orchestration in multi-agent environments."} | |
| {"authors":["Luo, XY","Fu, YL","Li, XL","Li, SB"],"type":"Journal Article","year":"2023","title":"Dynamic event-Based resilient consensus of networked lagrangian systems under DoS attacks","venue":"JOURNAL OF THE FRANKLIN INSTITUTE-ENGINEERING AND APPLIED MATHEMATICS","volume":"360","number":"12","pages":"9198-9215","doi":"10.1016/j.jfranklin.2022.09.046","wos_id":"WOS:001047157100001","abstract":"<p>In this paper, the dynamic event-based resilient consensus control of the multiple networked Euler-Lagrangian (E-L) systems under the Denial of Service (DoS) attacks is considered. Compared with linear cyber-physical systems, nonlinear networked E-L systems are more complex and closer to actual mechanical systems. For the situation where the topology is a strongly connected directed topology, a controller based on a dynamic event-trigger mechanism is designed to achieve consensus control for the networked E-L system in the absence of DoS attacks. Sufficient conditions are presented, which can guarantee the closed-loop system be stable. Then the resilient consensus problem of event-based controllers under energy-constrained DoS attacks is analyzed. The conditions related to the duration and frequency of DoS attacks are given. Zeno behavior is proved does not exist in the proposed control scheme. Finally, some numerical simulation results are given for verifying the theoretical results. & COPY; 2022 The Franklin Institute. Published by Elsevier Ltd. All rights reserved.</p>","id":"9ce965ee2bd9","score":8,"tags":["multi-agent","event-driven","resilient consensus","DoS attacks","networked systems"],"keep":true,"rationale":"The paper discusses event-based control in multi-agent systems, which is relevant for adaptive orchestration and safety in the context of networked systems."} | |
| {"authors":["Ma, ZK","Sun, QC","Matsumaru, T"],"type":"Journal Article","year":"2024","title":"Bidirectional Planning for Autonomous Driving Framework with Large Language Model","venue":"SENSORS","volume":"24","number":"20","doi":"10.3390/s24206723","wos_id":"WOS:001341419900001","abstract":"<p>Autonomous navigation systems often struggle in dynamic, complex environments due to challenges in safety, intent prediction, and strategic planning. Traditional methods are limited by rigid architectures and inadequate safety mechanisms, reducing adaptability to unpredictable scenarios. We propose SafeMod, a novel framework enhancing safety in autonomous driving by improving decision-making and scenario management. SafeMod features a bidirectional planning structure with two components: forward planning and backward planning. Forward planning predicts surrounding agents' behavior using text-based environment descriptions and reasoning via large language models, generating action predictions. These are embedded into a transformer-based planner that integrates text and image data to produce feasible driving trajectories. Backward planning refines these trajectories using policy and value functions learned through Actor-Critic-based reinforcement learning, selecting optimal actions based on probability distributions. Experiments on CARLA and nuScenes benchmarks demonstrate that SafeMod outperforms recent planning systems in both real-world and simulation testing, significantly improving safety and decision-making. This underscores SafeMod's potential to effectively integrate safety considerations and decision-making in autonomous driving.</p>","id":"1ddfed6f9b19","score":6,"tags":["autonomous driving","safety","multi-agent","planning","decision-making"],"keep":true,"rationale":"The paper discusses safety and decision-making in a multi-agent context, which is relevant to the orchestration of agents, though it focuses on autonomous driving rather than a general multi-agent orchestration framework."} | |
| {"authors":["Marandi, S","Hu, YS","Modarres, M"],"type":"Journal Article","year":"2025","title":"Complex System Diagnostics Using a Knowledge Graph-Informed and Large Language Model-Enhanced Framework","venue":"APPLIED SCIENCES-BASEL","volume":"15","number":"17","doi":"10.3390/app15179428","wos_id":"WOS:001569596400001","abstract":"<p>Featured Application This framework demonstrates how Large Language Models (LLMs) can be used to automate functional model construction and enable natural language-driven fault diagnostics in complex engineered systems. By integrating LLMs with a knowledge graph of an Auxiliary Feedwater system, the approach supports predictive maintenance and intelligent fault analysis. It applies to safety-critical domains such as nuclear power plants and other high-reliability industries.Abstract This paper presents a hybrid diagnostic framework that integrates Knowledge Graphs (KGs) with Large Language Models (LLMs) to support fault diagnosis in complex, high-reliability systems such as nuclear power plants. The framework is based on the Dynamic Master Logic (DML) model, which organizes system functions, components, and dependencies into a hierarchical KG for logic-based reasoning. LLMs act as high-level facilitators by automating the extraction of DML logic from unstructured technical documentation, linking functional models with language-based reasoning, and interpreting user queries in natural language. For diagnostic queries, the LLM agent selects and invokes predefined tools that perform upward or downward propagation in the KG using DML logic, while explanatory queries retrieve and contextualize relevant KG segments to generate user-friendly interpretations. This ensures that reasoning remains transparent and grounded in the system structure. This approach reduces the manual effort needed to construct functional models and enables natural language queries to deliver diagnostic insights. In a case study on an auxiliary feedwater system used in the nuclear pressurized water reactors, the framework achieved over 90 percent accuracy in model element extraction and consistently interpreted both diagnostic and explanatory queries. The results validate the effectiveness of LLMs in automating model construction and delivering explainable AI-assisted health monitoring.</p>","id":"37cacbe5603e","score":8,"tags":["Knowledge Graph","Large Language Models","Fault Diagnosis","Complex Systems","Safety-Critical Systems"],"keep":true,"rationale":"The paper discusses the integration of knowledge graphs and LLMs for fault diagnosis in safety-critical systems, which aligns with the themes of adaptive orchestration and safety in multi-agent systems."} | |
| {"authors":["Mullen , JF Jr","Goyal, P","Piramuthu, R","Johnston, M","Manocha, D","Ghanadan, R"],"type":"Journal Article","year":"2024","title":"\"Don't Forget to Put the Milk Back!\" Dataset for Enabling Embodied Agents to Detect Anomalous Situations","venue":"IEEE ROBOTICS AND AUTOMATION LETTERS","volume":"9","number":"10","pages":"9087-9094","doi":"10.1109/LRA.2024.3430129","wos_id":"WOS:001316210300010","abstract":"<p>Home robots intend to make their users lives easier. Our work moves toward more helpful home robots by enabling them to inform their users of dangerous or unsanitary anomalies in the home. Some examples of these anomalies include the user leaving their milk out, forgetting to turn off the stove, or leaving poison accessible to children. To enable home robots with these abilities, we have created a new dataset, which we call SafetyDetect. The SafetyDetect dataset consists of 1000 anomalous home scenes, each of which contains unsafe or unsanitary situations for an agent to detect. Our approach utilizes large language models (LLMs) alongside both a graph representation of the scene which encodes relationships between the objects in the scene. Our key insight is that this connected scene graph and the object relationships it encodes enables the LLM to better reason about the scene - especially as it relates to detecting dangerous or unsanitary situations. Our most promising approach utilizes GPT-4 and pursues a classification technique where object relations from the scene graph are classified as normal, dangerous, unsanitary, or dangerous for children. This method is able to correctly identify over 90% of anomalous scenarios in the SafetyDetect Dataset. Additionally, we conduct real world experiments on a ClearPath TurtleBot where we generate a scene graph from visuals of the real world scene, and run our approach with no modification. This setup resulted in little performance loss. The SafetyDetect Dataset and code will be released to the public upon this papers publication.</p>","id":"5fb950f4513c","score":8,"tags":["multi-agent","graph","safety","LLM","anomaly detection"],"keep":true,"rationale":"The paper discusses a dataset and methodology for enabling agents to detect anomalies in home environments using graph representations, which aligns with the themes of multi-agent orchestration and safety."} | |
| {"authors":["Nandanwar, A","Dhar, NK","Behera, L","Sinha, R"],"type":"Journal Article","year":"2023","title":"Near-optimal sliding mode control for multi-robot consensus under dynamic events","venue":"ADVANCED ROBOTICS","doi":"10.1080/01691864.2022.2155489","wos_id":"WOS:000921816200001","abstract":"<p>We propose a continuous-time design for finite-time consensus control for multi-robot system using event-based near-optimal sliding mode control. The system has a leader-follower framework prone to external bounded disturbance. The proposed design comprises of three parts: (i) formulation of control-affine dynamics, (ii) design a triggering condition for control updates that guarantee stability and consensus in the system, and (iii) design a near-optimal sliding mode control using neural-network based approximate dynamic programming. We derive a bound on inter-event time that guarantees admissibility of updated control input values. We finally validate the efficacy of proposed design through real-time experiments using three Pioneer P3-DX mobile robots (leader and two followers) and comparative analyses with other state-of-the-art approaches. The control updates of follower-1 and follower-2 robots are approximately 30.00% and 32.22%, respectively, that reduce the computational burden in multi-robot framework.</p>","id":"f91eeebc0bd5","score":8,"tags":["multi-agent","event-driven","control","consensus","robotics"],"keep":true,"rationale":"The paper discusses event-driven control for multi-robot systems, which is relevant to adaptive orchestration in multi-agent environments."} | |
| {"authors":["Nasri, M","Ginn, HL","Moallem, M"],"type":"Journal Article","year":"2021","title":"Agent-Based Coordinated Control of Power Electronic Converters in a Microgrid","venue":"ELECTRONICS","volume":"10","number":"9","doi":"10.3390/electronics10091031","wos_id":"WOS:000649988200001","abstract":"<p>This paper presents the implementation of an agent-based architecture suitable for the coordination of power electronic converters in stand-alone microgrids. To this end, a publish-subscribe agent architecture was utilized as a distributed microgrid control platform. Over a distributed hash table (DHT) searching overlay, the publish-subscribe architecture was identified based on a numerical analysis as a scalable agent-based technology for the distributed real-time coordination of power converters in microgrids. The developed framework was set up to deploy power-sharing distributed optimization algorithms while keeping a deterministic time period of a few tens of milliseconds for a system with tens of converters and when multiple events might happen concurrently. Several agents participate in supervisory control to regulate optimum power-sharing for the converters. To test the design, a notional shipboard system, including several converters, was used as a case study. Results of implementing the agent-based publish-subscribe control system using the Java Agent Development Framework (JADE) are presented.</p>","id":"5a1ec92a304d","score":8,"tags":["multi-agent","orchestration","event-driven","microgrid","publish-subscribe"],"keep":true,"rationale":"The paper discusses an agent-based architecture for coordinating power converters in microgrids, which aligns with multi-agent orchestration and event-driven systems."} | |
| {"authors":["Ngo, VT","Liu, YC"],"type":"Journal Article","year":"2021","title":"Event-Based Communication and Control for Task-Space Consensus of Networked Euler-Lagrange Systems","venue":"IEEE TRANSACTIONS ON CONTROL OF NETWORK SYSTEMS","volume":"8","number":"2","pages":"555-565","doi":"10.1109/TCNS.2021.3050126","wos_id":"WOS:000690440800006","abstract":"<p>In this article, we study the problems of event-based communication and actuation algorithms to mitigate unnecessary network burden and energy consumption, hence increasing the number and work time of robots/agents in a network of multiagent systems. By considering the general dynamic model of networked Euler-Lagrange systems and utilizing the adaptive control algorithm, we propose the following two event-triggered schemes: first, an event-based communication and, second, an event-based controller for a large number of agents to achieve consensus in the task space. Consider that the network connection is a directed spanning tree and the time-varying communication delays are bounded. Theoretical analyses of the proposed control algorithms for both leaderless and leader-follower (static leader) consensus are studied by employing the Lyapunov technique and function analysis. It revealed that the network achieved global stability and asymptotical convergence with avoidance of Zeno behavior. We experimented with a system of four robotic manipulators and performed numerical simulations for a networked mobile manipulator to demonstrate the efficiency and efficacy of the proposed consensus algorithms.</p>","id":"e43ac9bbf90c","score":9,"tags":["multi-agent systems","event-driven","consensus","control algorithms","networked systems"],"keep":true,"rationale":"The paper discusses event-based communication and control in multi-agent systems, which is highly relevant to the orchestration and adaptive event-driven aspects of the dissertation."} | |
| {"authors":["Ni, MZ","Wang, T","Leng, JW","Chen, C","Cheng, LL"],"type":"Journal Article","year":"2025","title":"A large language model-based manufacturing process planning approach under industry 5.0","venue":"INTERNATIONAL JOURNAL OF PRODUCTION RESEARCH","doi":"10.1080/00207543.2025.2469285","wos_id":"WOS:001432578100001","abstract":"<p>Industry 5.0 witnesses a new era where human intelligence and smart technology converge to redefine manufacturing. Amid this transformation, the ability to dynamically generate adaptable manufacturing processes is crucial for meeting the demands of personalised and flexible production. In order to achieve accurate manufacturing process planning, our research introduces LLM Adaptive Process Management (LLMAPM), a strategy that employs Large Language Models (LLMs) to transform user descriptions into structured manufacturing task flows, thereby enhancing the flexibility and responsiveness of production systems. LLMAPM adopts a three-phase methodology: task splitting, step generation, and holistic process synthesis. Beginning with informal user inputs, the system undergoes formal expansion before diving into granular step definitions. Subsequently, these elements are integrated to form a complete workflow. Finally, state machines are integrated to validate the logical accuracy and safety of the generated processes. Extensive experiments on a low-code industrial software platform are conducted to validate the effectiveness of the proposed study. The results indicate LLMAPM's capability to seamlessly coordinate manufacturing devices, confirming enhancements in workflow generation efficiency, deployment flexibility, and overall process accuracy.</p>","id":"caa7ea247a1b","score":4,"tags":["manufacturing","process planning","LLM","flexibility","workflow"],"keep":false,"rationale":"While the paper discusses adaptive processes and workflow generation, it focuses on manufacturing rather than multi-agent orchestration or event-driven systems."} | |
| {"authors":["Ni, Y","Jia, FL"],"type":"Journal Article","year":"2025","title":"A Scoping Review of AI-Driven Digital Interventions in Mental Health Care: Mapping Applications Across Screening, Support, Monitoring, Prevention, and Clinical Education","venue":"HEALTHCARE","volume":"13","number":"10","doi":"10.3390/healthcare13101205","wos_id":"WOS:001495618100001","abstract":"<p>Background/Objectives: Artificial intelligence (AI)-enabled digital interventions are increasingly used to expand access to mental health care. This PRISMA-ScR scoping review maps how AI technologies support mental health care across five phases: pre-treatment (screening), treatment (therapeutic support), post-treatment (monitoring), clinical education, and population-level prevention. Methods: We synthesized findings from 36 empirical studies published through January 2024 that implemented AI-driven digital tools, including large language models (LLMs), machine learning (ML) models, and conversational agents. Use cases include referral triage, remote patient monitoring, empathic communication enhancement, and AI-assisted psychotherapy delivered via chatbots and voice agents. Results: Across the 36 included studies, the most common AI modalities included chatbots, natural language processing tools, machine learning and deep learning models, and large language model-based agents. These technologies were predominantly used for support, monitoring, and self-management purposes rather than as standalone treatments. Reported benefits included reduced wait times, increased engagement, and improved symptom tracking. However, recurring challenges such as algorithmic bias, data privacy risks, and workflow integration barriers highlight the need for ethical design and human oversight. Conclusion: By introducing a four-pillar framework, this review offers a comprehensive overview of current applications and future directions in AI-augmented mental health care. It aims to guide researchers, clinicians, and policymakers in developing safe, effective, and equitable digital mental health interventions.</p>","id":"8d85e66a9b55","score":4,"tags":["AI","mental health","digital interventions","chatbots","machine learning"],"keep":false,"rationale":"While the paper discusses AI-driven interventions and chatbots, it does not focus on multi-agent orchestration or the specific infrastructure aspects relevant to the dissertation topic."} | |
| {"authors":["Niazi, SK"],"type":"Journal Article","year":"2025","title":"Artificial Intelligence in Small-Molecule Drug Discovery: A Critical Review of Methods, Applications, and Real-World Outcomes","venue":"PHARMACEUTICALS","volume":"18","number":"9","doi":"10.3390/ph18091271","wos_id":"WOS:001579926600001","abstract":"<p>Artificial intelligence (AI) is emerging as a valuable complementary tool in small-molecule drug discovery, augmenting traditional methodologies rather than replacing them. This review examines the evolution of AI from early rule-based systems to advanced deep learning, generative models, diffusion models, and autonomous agentic AI systems, highlighting their applications in target identification, hit discovery, lead optimization, and safety prediction. We present both successes and failures to provide a balanced perspective. Notable achievements include baricitinib (BenevolentAI/Eli Lilly, an existing drug repurposed through AI-assisted analysis for COVID-19 and rheumatoid arthritis), halicin (MIT, preclinical antibiotic), DSP-1181 (Exscientia, discontinued after Phase I), and ISM001-055/rentosertib (Insilico Medicine, positive Phase IIa results). However, several AI-assisted compounds have also faced challenges in clinical development. DSP-1181 was discontinued after Phase I, despite a favorable safety profile, highlighting that the acceleration of discovery timelines by AI does not guarantee clinical success. Despite progress, challenges such as data quality, model interpretability, regulatory hurdles, and ethical concerns persist. We provide practical insights for integrating AI into drug discovery workflows, emphasizing hybrid human-AI approaches and the emergence of agentic AI systems that can autonomously navigate discovery pipelines. A critical evaluation of current limitations and future opportunities reveals that while AI offers significant potential as a complementary technology, realistic expectations and careful implementation are crucial for delivering innovative therapeutics.</p>","id":"959a87492175","score":4,"tags":["AI","drug discovery","agentic systems","workflow integration"],"keep":false,"rationale":"While the paper discusses agentic AI systems, its focus on drug discovery does not align closely with the core themes of multi-agent orchestration and event-driven infrastructure."} | |
| {"authors":["Omeñaca, AT","Rocabruna, EL","Sloan, J","Catta-Preta, M","Picó, JFI","Alvarez, JCA","Solis, TA","Gil, EL","Vinaixa, XS","Villegas, DV","Garcia, RR","Feijoo, CR","Fierro, JM","Genis, BB"],"type":"Journal Article","year":"2025","title":"Leave as Fast as You Can: Using Generative AI to Automate and Accelerate Hospital Discharge Reports","venue":"COMPUTERS","volume":"14","number":"6","doi":"10.3390/computers14060210","wos_id":"WOS:001514764900001","abstract":"<p>Clinical documentation, particularly the hospital discharge report (HDR), is essential for ensuring continuity of care, yet its preparation is time-consuming and places a considerable clinical and administrative burden on healthcare professionals. Recent advancements in Generative Artificial Intelligence (GenAI) and the use of prompt engineering in large language models (LLMs) offer opportunities to automate parts of this process, improving efficiency and documentation quality while reducing administrative workload. This study aims to design a digital system based on LLMs capable of automatically generating HDRs using information from clinical course notes and emergency care reports. The system was developed through iterative cycles, integrating various instruction flows and evaluating five different LLMs combined with prompt engineering strategies and agent-based architectures. Throughout the development, more than 60 discharge reports were generated and assessed, leading to continuous system refinement. In the production phase, 40 pneumology discharge reports were produced, receiving positive feedback from physicians, with an average score of 2.9 out of 4, indicating the system's usefulness, with only minor edits needed in most cases. The ongoing expansion of the system to additional services and its integration within a hospital electronic system highlights the potential of LLMs, when combined with effective prompt engineering and agent-based architectures, to generate high-quality medical content and provide meaningful support to healthcare professionals. Hospital discharge reports (HDRs) are pivotal for continuity of care but consume substantial clinician time. Generative AI systems based on large language models (LLMs) could streamline this process, provided they deliver accurate, multilingual, and workflow-compatible outputs. We pursued a three-stage, design-science approach. Proof-of-concept: five state-of-the-art LLMs were benchmarked with multi-agent prompting to produce sample HDRs and define the optimal agent structure. Prototype: 60 HDRs spanning six specialties were generated and compared with clinician originals using ROUGE with average scores compatible with specialized news summarizing models in Spanish and Catalan (lower scores). A qualitative audit of 27 HDR pairs showed recurrent divergences in medication dose (56%) and social context (52%). Pilot deployment: The AI-HDR service was embedded in the hospital's electronic health record. In the pilot, 47 HDRs were autogenerated in real-world settings and reviewed by attending physicians. Missing information and factual errors were flagged in 53% and 47% of drafts, respectively, while written assessments diminished the importance of these errors. An LLM-driven, agent-orchestrated pipeline can safely draft real-world HDRs, cutting administrative overhead while achieving clinician-acceptable quality, not without errors that require human supervision. Future work should refine specialty-specific prompts to curb omissions, add temporal consistency checks to prevent outdated data propagation, and validate time savings and clinical impact in multi-center trials.</p>","id":"59c02469d2b7","score":6,"tags":["Generative AI","Hospital Discharge Reports","Multi-Agent Systems","LLMs","Automation"],"keep":true,"rationale":"The paper discusses the use of agent-based architectures and LLMs, which are relevant to the orchestration of multi-agent systems, although it focuses on a specific application in healthcare."} | |
| {"authors":["Pan, WH","Lin, BB","Wang, YF","Yu, ZX","Zhao, XK","He, XF","Ye, JP"],"type":"Journal Article","year":"2025","title":"Cooperative Driving at Multiple Unsignalized Intersections in Fully Autonomous Driving Scenarios","venue":"IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS","doi":"10.1109/TITS.2025.3615073","wos_id":"WOS:001587366200001","abstract":"<p>The decision-making process for connected and autonomous vehicles (CAVs) at unsignalized intersections is a critical and challenging problem. Previous methods predominantly concentrate on optimizing passage strategies for individual intersections in isolation. However, they often neglect global traffic conditions and task priorities in closed, multi-intersection transportation scenarios, leading to localized congestion. In this work, we propose a method that aims to optimize the passing order of intersections from a global and long-term perspective to enhance overall transportation efficiency. Specifically, we model the coordination of multiple unsignalized intersections as a multi-agent sequential decision problem and solve it through a two-stage method. In the planning stage, we construct fully connected undirected graphs based on vehicle conflict relationships and use the multi-agent proximal policy optimization (MAPPO) algorithm to learn the global priorities. In the scheduling stage, the local vehicle scheduling is formalized as a multi-objective optimization problem. The learned global priorities are soft constraints, while a hybrid filtered beam search determines safe and efficient CAV passing orders. Extensive offline experiments and online tests on real-world and synthetic datasets demonstrate that our proposed method outperforms state-of-the-art approaches in minimizing congestion and enhancing transportation efficiency.</p>","id":"cbc994c88562","score":9,"tags":["multi-agent","orchestration","graph","safety","adaptive"],"keep":true,"rationale":"The paper addresses multi-agent decision-making in a transportation context, utilizing graph-based models and optimization techniques relevant to safe orchestration."} | |
| {"authors":["Park, JH","Madisetti, VK"],"type":"Journal Article","year":"2025","title":"CAPRI: A Context-Aware Privacy Framework for Multi-Agent Generative AI Applications","venue":"IEEE ACCESS","volume":"13","pages":"43168-43177","doi":"10.1109/ACCESS.2025.3549312","wos_id":"WOS:001445065100032","abstract":"<p>While the swift advancement of cloud-based Large Language Models (LLMs) has significantly increased the efficiency and automation in business processes, it has also introduced considerable privacy concerns regarding Personally Identifiable Information (PII) and other protected data in multimodal forms, such as text, video, or images, being exported, potentially insecurely, outside the corporate environments. Although traditional anonymization-based techniques can alleviate these risks in offline applications, such as summarization or classification, incorporating it into online LLM workflows poses substantial challenges, particularly when these workflows encompass real-time transactions involving multiple stakeholders, as commonly observed in multi-agent generative AI applications. This study explores these challenges and proposes novel context-aware privacy frameworks and methods to address these issues. We employ a local privacy-focused gatekeeper LLM to contextually pseudonymize PII and assign unique identifiers as part of a new mapping process, thereby facilitating re-identification in real-time operations while safeguarding privacy when interacting with cloud-based LLMs. Our proposed methodologies and frameworks adeptly integrate privacy considerations into LLM and LLM Agent workflows, preserving both privacy and data utility while maintaining operational efficiency and utility comparable to non-anonymized generative AI processes.</p>","id":"488705bbaf7c","score":7,"tags":["multi-agent","privacy","LLM","workflow","context-aware"],"keep":true,"rationale":"The paper discusses a context-aware privacy framework for multi-agent applications, which is relevant to the orchestration and safety aspects of multi-agent systems."} | |
| {"authors":["Queffelec, A","Sankur, O","Schwarzentruber, F"],"type":"Journal Article","year":"2023","title":"Complexity of planning for connected agents in a partially known environment","venue":"THEORETICAL COMPUTER SCIENCE","volume":"941","pages":"202-220","doi":"10.1016/j.tcs.2022.11.015","wos_id":"WOS:000906113000016","abstract":"<p>The Connected Multi-Agent Path Finding (CMAPF) problem asks for a plan to move a group of agents in a graph while respecting a connectivity constraint. We study a generalization of CMAPF in which the graph is not entirely known in advance, but is discovered by the agents during their mission. We present a framework introducing this notion and study the problem of searching for a strategy to reach a configuration in this setting. We prove the problem to be PSPACE-complete when requiring all agents to be connected at all times, and NEXPTIME-hard in the decentralized case.(c) 2022 Elsevier B.V. All rights reserved.</p>","id":"9b73de391da3","score":9,"tags":["multi-agent","graph","planning","connectivity","complexity"],"keep":true,"rationale":"The paper addresses the planning complexity for connected agents in a graph, which is highly relevant to the orchestration of multi-agent systems."} | |
| {"authors":["Rikos, AI","Hadjicostis, CN","Johansson, KH"],"type":"Journal Article","year":"2022","title":"Non-oscillating quantized average consensus over dynamic directed topologies","venue":"AUTOMATICA","volume":"146","doi":"10.1016/j.automatica.2022.110621","wos_id":"WOS:000870698100013","abstract":"<p>In this paper we study the distributed average consensus problem in multi-agent systems with dynamically-changing directed communication links that are subject to quantized information flow. We present and analyze a distributed averaging algorithm which operates exclusively with quantized values (i.e., the information stored, processed and exchanged between neighboring agents is subject to deterministic uniform quantization) and relies on event-driven updates (e.g., to reduce energy con-sumption, communication bandwidth, network congestion, and/or processor usage). We characterize the properties of the proposed distributed algorithm over dynamic directed communication topologies subject to some connectivity conditions and we show that its execution allows each agent to reach, in finite time, a fixed state that is equal (within one quantization level) to the average of the initial states. The main idea of the proposed algorithm is that each agent (i) models its initial state as two quantized fractions which have numerators equal to the agent's initial state and denominators equal to one, and (ii) transmits one fraction randomly while it keeps the other stored. Then, every time an agent receives one or more fractions, it averages their numerators with the numerator of the fraction it stored, and then transmits them to randomly selected out-neighbors. Finally, we provide examples to illustrate the operation, performance, and potential advantages of the proposed algorithm. We compare against various quantized average consensus algorithms and show that our algorithm's convergence speed is among the fastest in the current literature.(c) 2022 Published by Elsevier Ltd.</p>","id":"25c7abcb8fc3","score":9,"tags":["multi-agent systems","event-driven","graph-based","average consensus","dynamic topologies"],"keep":true,"rationale":"The paper addresses distributed average consensus in multi-agent systems with event-driven updates, which is highly relevant to the orchestration of agents in dynamic environments."} | |
| {"authors":["Rikos, AI","Hadjicostis, CN","Johansson, KH"],"type":"Journal Article","year":"2024","title":"Finite time quantized average consensus with transmission stopping guarantees and no quantization error","venue":"AUTOMATICA","volume":"163","doi":"10.1016/j.automatica.2024.111522","wos_id":"WOS:001180494100001","abstract":"<p>Networked control systems, which are composed of spatially distributed sensors and actuators that communicate through wireless networks, are emerging as a fundamental infrastructure technology in 5G and IoT technologies. In order to increase flexibility and reduce deployment and maintenance costs, their operation needs to guarantee (i) efficient communication between nodes and (ii) preservation of available energy. Motivated by these requirements, we present and analyze a novel distributed average consensus algorithm, which (i) operates exclusively on quantized values (in order to guarantee efficient communication and data storage), (ii) relies on event-driven updates (in order to reduce energy consumption, communication bandwidth, network congestion, and/or processor usage), and (iii) allows each node to cease transmissions once the exact average of the initial quantized values has been reached (in order to preserve its stored energy). We characterize the properties of the proposed algorithm and show that its execution, on any time-invariant and strongly connected digraph, allows all nodes to reach in finite time a common consensus value that is equal to the exact average (represented as the ratio of two quantized values). Then, we present upper bounds on (i) the number of transmissions and computations each node has to perform during the execution of the algorithm, and (ii) the memory and energy requirements of each node in order for the algorithm to be executed. Finally, we provide examples that demonstrate the operation, performance, and potential advantages of our proposed algorithm. (c) 2024 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).</p>","id":"8b223b5e7738","score":8,"tags":["multi-agent","event-driven","graph","consensus","energy efficiency"],"keep":true,"rationale":"The paper discusses a distributed average consensus algorithm that operates on quantized values and uses event-driven updates, which aligns well with the themes of adaptive event-driven systems and multi-agent orchestration."} | |
| {"authors":["Rokhforoz, P","Kebriaei, H","Ahmadabadi, MN"],"type":"Journal Article","year":"2021","title":"Large-scale dynamic system optimization using dual decomposition method with approximate dynamic programming","venue":"SYSTEMS & CONTROL LETTERS","volume":"150","doi":"10.1016/j.sysconle.2021.104894","wos_id":"WOS:000637970700003","abstract":"<p>In this paper, multi-agent dynamic optimization with a coupling constraint is studied. The aim is to minimize a strongly convex social cost function, by considering a linear stochastic dynamics for each agent and also coupling constraints among the agents. In order to handle the coupling constraint and also, to avoid high computational cost imposed by a centralized method for large scale systems, the dual decomposition method is used to decompose the problem into multiple individual sub-problems, while the dual variable is adjusted by a coordinator. Nevertheless, since each sub-problem is not a linear-quadratic (LQ) optimal control problem, and hence its closed-form solution does not exist, approximate dynamic programming (ADP) is utilized to solve the sub-problems. The main contribution of the paper is to propose an algorithm by considering the interrelated iterations of dual variable adjustment and ADP, and to prove the convergence of the algorithm to the global optimal solution of the social cost function. Additionally, the implementation of the proposed algorithm using a neural network is presented. Also, the computational advantage of the proposed algorithm in comparison with other bench-marking methods is discussed in simulation results. (C) 2021 Elsevier B.V. All rights reserved.</p>","id":"9076afadc400","score":8,"tags":["multi-agent","dynamic optimization","dual decomposition","approximate dynamic programming"],"keep":true,"rationale":"The paper addresses multi-agent dynamic optimization, which is relevant for understanding orchestration in agent-based systems, particularly in the context of coupling constraints and optimization methods."} | |
| {"authors":["Shi, YX","Hu, QL"],"type":"Journal Article","year":"2021","title":"Observer-Based Spacecraft Formation Coordinated Control via a Unified Event-Triggered Communication","venue":"IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS","volume":"57","number":"5","pages":"3307-3319","doi":"10.1109/TAES.2021.3074201","wos_id":"WOS:000704826600055","abstract":"<p>This article addresses the formation tracking control problem for multiple spacecraft systems subject to limited communication resources and external disturbances. Considering that only a subset of the follower spacecraft can have access to the motion states of the dynamic leader, an event-based distributed observer is first developed to reconstruct the leader's information. Subsequently, to achieve the accompanying flight of the follower spacecraft around the leader with a desired formation configuration, for each follower spacecraft, a distributed event-triggered coordinated controller is proposed. In particular, by embedding simultaneously the position information and observer output into the triggering function, a novel unified event-triggered mechanism is designed to schedule information transmission in multiple spacecraft systems. The salient characters of the distributed coordinated control scheme are twofold: unnecessary occupancy of communication resources can be avoided significantly; and asymptotic stability of the whole closed-loop system is guaranteed without resorting to the separation principle. Finally, numerical simulations are carried out to illustrate the efficiency of the theoretical results.</p>","id":"fb3941fa6040","score":8,"tags":["multi-agent","event-driven","formation control","communication","safety"],"keep":true,"rationale":"The paper discusses event-triggered communication and coordinated control in multi-agent systems, which is relevant to adaptive orchestration in agent-based infrastructures."} | |
| {"authors":["Shi, YX","Hu, QL"],"type":"Journal Article","year":"2022","title":"Event-Driven Connectivity-Preserving Coordinated Control for Multiple Spacecraft Systems With a Distance-Dependent Dynamic Graph","venue":"IEEE TRANSACTIONS ON CYBERNETICS","volume":"52","number":"11","pages":"12551-12560","doi":"10.1109/TCYB.2021.3072139","wos_id":"WOS:000732168200001","abstract":"<p>This article considers the connectivity preservation coordinated control problem for multiple spacecraft systems subject to limited communication resources and sensing capability. By constructing a novel bump function, a distance-dependent dynamic communication network model is first presented, which characterizes the interaction strength as a nonlinear smooth function varying with the relative distance of spacecraft continuously. Subsequently, based on an edge-tension potential function, a distributed event-driven coordinated control scheme is proposed to achieve formation consensus, while ensuring that adjacent spacecraft is always within the allowable connectivity range. Meanwhile, to avoid redundant data transmissions, a hybrid dynamic event-triggered mechanism with maximum triggering interval is developed to schedule the communication frequency among spacecraft. It is proven that the onboard communication resources occupation can be reduced significantly and the Zeno phenomenon is strictly excluded. Finally, the efficiency of the proposed method for, as an example, four-spacecraft formation system is substantiated.</p>","id":"98479ee94533","score":9,"tags":["multi-agent","event-driven","graph","safety"],"keep":true,"rationale":"The paper discusses event-driven control in multi-agent systems with a focus on connectivity and communication, which aligns closely with the themes of adaptive orchestration and graph-based infrastructure."} | |
| {"authors":["Stein, S","Pilgermann, M","Weber, S","Sedlmayr, M"],"type":"Journal Article","year":"2025","title":"Leveraging MDS2 and SBOM data for LLM-assisted vulnerability analysis of medical devices","venue":"COMPUTATIONAL AND STRUCTURAL BIOTECHNOLOGY JOURNAL","volume":"28","pages":"267-280","doi":"10.1016/j.csbj.2025.07.012","wos_id":"WOS:001547399200001","abstract":"<p>This study investigated the use of a semi-automated, Retrieval-Augmented Generation (RAG)-based multi-agent architecture to analyze security-relevant data and assemble specialized exploitation paths targeting medical devices. The input dataset comprised device-specific sources, namely, the Manufacturer Disclosure Statement for Medical Device Security (MDS2) documents and Software Bills of Materials (SBOMs), enriched with public vulnerability databases, including Common Vulnerabilities and Exposures (CVE), Known Exploited Vulnerabilities (KEV), and Metasploit exploit records. The objective was to assess whether a modular, Large Language Model (LLM)-driven agent system could autonomously correlate device metadata with known vulnerabilities and existing exploit information to support structured threat modeling. The architecture follows a static RAG design based on predefined prompts and fixed retrieval logic, without autonomous agent planning or dynamic query adaptation. The developed Vulnerability Intelligence for Threat Analysis in Medical Security (VITAMedSec) system operates under human-prompted supervision and successfully synthesizes actionable insights and exploitation paths without requiring manual step-by-step input during execution. Although technically coherent results were obtained under controlled conditions, real-world validation remains a critical avenue for future research. This study further discusses the dual-use implications of such an agent-based framework, its relevance to patient safety in medical device cybersecurity, and the broader applicability of the proposed architecture to other critical infrastructure sectors. These findings emphasize both the technical potential and ethical responsibility for applying semi-automated AI workflows in medical cybersecurity contexts.</p>","id":"8bd5ab81463f","score":6,"tags":["multi-agent","vulnerability analysis","medical devices","LLM","cybersecurity"],"keep":true,"rationale":"The paper discusses a multi-agent architecture for vulnerability analysis, which aligns with the themes of agent orchestration and safety in critical systems, though it focuses on medical devices rather than general multi-agent orchestration."} | |
| {"authors":["Sun, JY","Tan, ZL","Liu, S","Zhang, HG","Chuo, WY"],"type":"Journal Article","year":"2023","title":"Fully Distributed Event-Driven Coordination With Actuator Faults","venue":"IEEE TRANSACTIONS ON CYBERNETICS","volume":"53","number":"10","pages":"6456-6464","doi":"10.1109/TCYB.2022.3198499","wos_id":"WOS:000849228800001","abstract":"<p>This article investigates the event-driven fault-tolerance (ETFT) consensus problem for general linear multiagent systems (MASs) with partial loss of effectiveness (PLOE) and bias faults of actuators in leader-follower networks. Each agent's controller is only updated relatively infrequently at its event moments. A desirable feature of this article is that the proposed event-driven algorithm is fully distributed also independent of global information and additive fault boundaries. Based on this, a consensus error prediction model is used to avoid the limitation that each agent needs to monitor its neighbors' state under event-driven conditions continuously. We further excluded the Zeno behavior by proving that any adjacent event interval for each agent is greater than zero. The simulations verify our results.</p>","id":"5809b2b1b587","score":9,"tags":["multi-agent systems","event-driven","fault tolerance","distributed systems"],"keep":true,"rationale":"The paper addresses event-driven coordination in multi-agent systems with a focus on fault tolerance, which is highly relevant to the dissertation's theme of adaptive event-driven orchestration."} | |
| {"authors":["Sun, Y","Liu, XK"],"type":"Journal Article","year":"2025","title":"Research and Application of a Multi-Agent-Based Intelligent Mine Gas State Decision-Making System","venue":"APPLIED SCIENCES-BASEL","volume":"15","number":"2","doi":"10.3390/app15020968","wos_id":"WOS:001403995500001","abstract":"<p>To address the issues of low efficiency in manual processing and lack of accuracy in judgment within traditional mine gas safety inspections, this paper designs and implements the Intelligent Mine Gas State Decision-Making System based on large language models (LLMs) and a multi-agent system. The system aims to enhance the accuracy of gas over-limit alarms and improve the efficiency of generating judgment reports. The system integrates the reasoning capabilities of LLMs and optimizes task allocation and execution efficiency of agents through the study of the hybrid multi-agent orchestration algorithm. Furthermore, the system establishes a comprehensive gas risk assessment knowledge base, encompassing historical alarm data, real-time monitoring data, alarm judgment criteria, treatment methods, and relevant policies and regulations. Additionally, the system incorporates several technologies, including retrieval-augmented generation based on human feedback mechanisms, tool management, prompt engineering, and asynchronous processing, which further enhance the application performance of the LLM in the gas status judgment system. Experimental results indicate that the system effectively improves the efficiency of gas alarm processing and the quality of judgment reports in coal mines, providing solid technical support for accident prevention and management in mining operations.</p>","id":"19fd53f86b2d","score":8,"tags":["multi-agent systems","orchestration","safety","decision-making","LLMs"],"keep":true,"rationale":"The paper discusses a multi-agent system for decision-making in gas safety inspections, which aligns well with the themes of orchestration and safety in the dissertation."} | |
| {"authors":["Testa, A","Carnevale, G","Notarstefano, G"],"type":"Journal Article","year":"2025","title":"A Tutorial on Distributed Optimization for Cooperative Robotics: From Setups and Algorithms to Toolboxes and Research Directions","venue":"PROCEEDINGS OF THE IEEE","volume":"113","number":"1","pages":"40-65","doi":"10.1109/JPROC.2025.3557698","wos_id":"WOS:001480236900001","abstract":"<p>Several interesting problems in multirobot systems can be cast in the framework of distributed optimization. Examples include multirobot task allocation, vehicle routing, target protection, and surveillance. While the theoretical analysis of distributed optimization algorithms has received significant attention, its application to cooperative robotics has not been investigated in detail. In this article, we show how notable scenarios in cooperative robotics can be addressed by suitable distributed optimization setups. Specifically, after a brief introduction on the widely investigated consensus optimization (most suited for data analytics) and on the partition-based setup (matching the graph structure in the optimization), we focus on two distributed settings modeling several scenarios in cooperative robotics, i.e., the so-called constraint-coupled and aggregative optimization frameworks. For each one, we consider use-case applications, and we discuss tailored distributed algorithms with their convergence properties. Then, we revise state-of-the-art toolboxes allowing for the implementation of distributed schemes on real networks of robots without central coordinators. For each use case, we discuss its implementation in these toolboxes and provide simulations and real experiments on networks of heterogeneous robots.</p>","id":"d97929a3650a","score":8,"tags":["multi-agent","cooperative robotics","distributed optimization","graph-based","event-driven"],"keep":true,"rationale":"The paper discusses distributed optimization in cooperative robotics, which is relevant to multi-agent orchestration and graph-based frameworks."} | |
| {"authors":["Venugopal, VK","Kumar, A","Tan, MO","Szarf, G"],"type":"Journal Article","year":"2025","title":"Curriculum check, 2025-equipping radiology residents for AI challenges of tomorrow","venue":"ABDOMINAL RADIOLOGY","doi":"10.1007/s00261-025-05016-5","wos_id":"WOS:001504385600001","abstract":"<p>The exponential rise in the artificial intelligence (AI) tools for medical imaging is profoundly impacting the practice of radiology. With over 1000 FDA-cleared AI algorithms now approved for clinical use-many of them designed for radiologic tasks-the responsibility lies with training institutions to ensure that radiology residents are equipped not only to use AI systems, but to critically evaluate, monitor, respond to their output in a safe, ethical manner. This review proposes a comprehensive framework to integrate AI into radiology residency curricula, targeting both essential competencies required of all residents, optional advanced skills for those interested in research or AI development. Core educational strategies include structured didactic instruction, hands-on lab exposure to commercial AI tools, case-based discussions, simulation-based clinical pathways, teaching residents how to interpret model cards, regulatory documentation. Clinical examples such as stroke triage, Urinary tract calculi detection, AI-CAD in mammography, false-positive detection are used to anchor theory in practice. The article also addresses critical domains of AI governance: model transparency, ethical dilemmas, algorithmic bias, the role of residents in human-in-the-loop oversight systems. It outlines mentorship, faculty development strategies to build institutional readiness, proposes a roadmap to future-proof radiology education. This includes exposure to foundation models, vision-language systems, multi-agent workflows, global best practices in post-deployment AI monitoring. This pragmatic framework aims to serve as a guide for residency programs adapting to the next era of radiology practice</p>","id":"4b4e5efcef4f","score":4,"tags":["AI","radiology","education","multi-agent","safety"],"keep":false,"rationale":"While the paper discusses AI in radiology and touches on multi-agent workflows, it primarily focuses on educational frameworks rather than the orchestration or infrastructure aspects relevant to the dissertation topic."} | |
| {"authors":["Waghchoure, MR","Patel, JK","Sanghai, N","Kanoun, S","John, RT","Gupta, G","Deshpande, BN","Dorle, A"],"type":"Journal Article","year":"2022","title":"A Receding Horizon Autopilot for the Two-Lane Highway Automated Driving Application through Synergy between the Robust Behavior Planner and the Advanced Driver Assistance Features","venue":"SAE INTERNATIONAL JOURNAL OF CONNECTED AND AUTOMATED VEHICLES","volume":"5","number":"3","pages":"271-292","doi":"10.4271/12-05-03-0022","wos_id":"WOS:001274888300006","abstract":"<p>Safety is always a crucial aspect of developing autonomous systems, and the motivation behind this project comes from the need to address the traffic crashes occurring globally on a daily basis. The present work studies the coexistence of the novel rule-based behavioral planning framework with the five key advanced driver assistance system (ADAS) features as proposed in this article to fulfill the safety requirements and enhance the comfort of the driver/passengers to achieve a receding-horizon autopilot. This architecture utilizes data from the sensor fusion and the prediction module for the prediction time horizon of 2 s iteratively, which is continuously moving forward (hence, the receding horizon), and helps the behavior planner understand the intent of other vehicles on the road in advance. Further, that information helps the behavior planner make an appropriate decision regarding the activation of specific ADAS features to drive safely on the highway, and that decision is being updated with every iteration or after 0.01 s. The driver assistance features are well equipped to deal with any eventuality on the road with proper guidance from the behavior planner Currently, there exist local, global, and behavior planners for planning the target trajectory of the ego vehicle (the vehicle that is comprised of the sensors that perceive the environment around it and which needs to operate with the intended level of autonomy) in order to deliver a safe and comfortable drive. Here, the goal of the system is for each ADAS feature to act independently and their synergy with the behavior planner leads to automated driving on the two-lane highway without the need for a global planner to guide it toward the goal. A finite-state machine consisting of a state flow model is used to switch between various driving modes based on the information from the behavior planner and the autopilot models. The behavioral planning framework incorporates a cost function library to determine the best set of ADAS features for the ego vehicle based on the lowest cost and its interaction with other actors in a complex and stochastic environment. The cost function-based algorithm ensures that the vehicle follows the traffic rules, safety, and comfort criteria without compromising performance and thus increases the robustness of the ADAS features leading to the autopilot capabilities for the two-lane one-way highway driving applications. The functionality of the behavior planner framework has been validated by incorporating the model-in-loop (MIL) testing method. The automated driving toolbox in MATLAB was used to perform MIL testing by creating an appropriate number of scenarios.</p>","id":"c469605478a1","score":6,"tags":["autonomous systems","behavior planning","ADAS","safety","multi-agent"],"keep":true,"rationale":"The paper discusses safety in autonomous systems and behavior planning, which is relevant to multi-agent orchestration and safety aspects of the dissertation."} | |
| {"authors":["Wang, H","Shan, JJ"],"type":"Journal Article","year":"2025","title":"Adaptive Dynamic Event-Based Robust Control for Multiple Networked Euler-Lagrange Systems","venue":"IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS","volume":"21","number":"9","pages":"6712-6722","doi":"10.1109/TII.2025.3567383","wos_id":"WOS:001494231300001","abstract":"<p>This article develops an event-based adaptive robust control scheme for multiple networked Euler-Lagrange systems with a dynamic leader, addressing some key challenges such as parameter uncertainties, unknown perturbations, inherent nonlinearities, and limited resources, for the practical applications of networked robotics and autonomous systems. To reduce the communication network burden and the computational resources consumption, an adaptive dynamic triggering strategy is developed. In addition, to estimate the inaccurate states, a nested adaptive sliding-mode estimator is proposed. Then, a fully distributed adaptive dynamic event-based time-varying sliding-mode control strategy is developed based on the designed triggering scheme and estimator, without requiring any global information. This strategy reduces the effect of large initial errors on the varying gain during adaptation, and compensates for the influences of inherent nonlinearities, unknown external perturbations, and parameter uncertainties, making it feasible for practical implementation. Moreover, Lyapunov stability theory is used to guarantee the asymptotic convergence of the closed-loop networked systems. Finally, hardware experiments are conducted using multiple quadrotors to validate the effectiveness of the proposed control scheme in multiagent coordination tasks.</p>","id":"6b500584d5f5","score":8,"tags":["multi-agent","event-driven","control systems","networked systems"],"keep":true,"rationale":"The paper discusses adaptive event-based control for networked systems, which is relevant to multi-agent orchestration and event-driven architectures."} | |
| {"authors":["Wang, J","He, GC","Kantaros, Y"],"type":"Journal Article","year":"2025","title":"Probabilistically Correct Language-Based Multi-Robot Planning Using Conformal Prediction","venue":"IEEE ROBOTICS AND AUTOMATION LETTERS","volume":"10","number":"1","pages":"160-167","doi":"10.1109/LRA.2024.3504233","wos_id":"WOS:001367266400014","abstract":"<p>This paper addresses task planning problems for language-instructed robot teams. Tasks are expressed in natural language (NL), requiring the robots to apply their skills at various locations and semantic objects. Several recent works have addressed similar planning problems by leveraging pre-trained Large Language Models (LLMs) to design effective multi-robot plans. However, these approaches lack performance guarantees. To address this challenge, we introduce a new distributed LLM-based planner, called S-ATLAS for Safe plAnning for Teams of Language-instructed AgentS, that can achieve user-defined mission success rates. This is accomplished by leveraging conformal prediction (CP), a distribution-free uncertainty quantification tool. CP allows the proposed multi-robot planner to reason about its inherent uncertainty, due to imperfections of LLMs, in a distributed fashion, enabling robots to make local decisions when they are sufficiently confident and seek help otherwise. We show, both theoretically and empirically, that the proposed planner can achieve user-specified task success rates, assuming successful plan execution, while minimizing the average number of help requests. We provide comparative experiments against related works showing that our method is significantly more computational efficient and achieves lower help rates.</p>","id":"534c344e94ff","score":9,"tags":["multi-agent","LLM","planning","safety","distributed systems"],"keep":true,"rationale":"The paper focuses on multi-robot planning using language models and emphasizes safety and performance guarantees, aligning well with the themes of adaptive orchestration and agent-based systems."} | |
| {"authors":["Wang, LQ","Dong, JX"],"type":"Journal Article","year":"2023","title":"Reset Event-Triggered Adaptive Fuzzy Consensus for Nonlinear Fractional-Order Multiagent Systems With Actuator Faults","venue":"IEEE TRANSACTIONS ON CYBERNETICS","volume":"53","number":"3","pages":"1868-1879","doi":"10.1109/TCYB.2022.3163528","wos_id":"WOS:000785844900001","abstract":"<p>This article studies the problem of event-triggered adaptive fault-tolerant fuzzy output feedback consensus tracking control for nonlinear fractional-order multiagent systems with actuator failures under a directed graph. Considering the fact that the actual system works near the equilibrium point most of the time, a novel dynamic event-triggering strategy with the reset mechanism is proposed, where the dynamic threshold can be actively adjusted according to the preset conditions, so that the resource utilization can be further reduced. Based on an improved event-based consensus error, the state estimator about the derivative of reference trajectory and the adaptive law about the information of graph are constructed, which makes distributed consensus tracking control achieved without obtaining global information. Then, by introducing two adaptive compensating terms to deal with actuator failures and event-triggered measurement errors, it is shown in the sense of fractional-order stability criterion that tracking errors can converge to a compact set even if the fault parameters and modes are completely unknown. Finally, the correctness of the presented method is verified by a simulation example.</p>","id":"09be7aa75e2b","score":9,"tags":["multi-agent systems","event-driven","graph theory","adaptive control","fault tolerance"],"keep":true,"rationale":"The paper addresses event-triggered adaptive control in multi-agent systems, which is highly relevant to the orchestration and adaptive aspects of the dissertation."} | |
| {"authors":["Wang, LQ","Dong, JX"],"type":"Journal Article","year":"2023","title":"Reset Event-Triggered Adaptive Fuzzy Consensus for Nonlinear Fractional-Order Multiagent Systems With Actuator Faults","venue":"IEEE TRANSACTIONS ON CYBERNETICS","volume":"53","number":"3","pages":"1868-1879","doi":"10.1109/TCYB.2022.3163528","wos_id":"WOS:000785844900001","abstract":"<p>This article studies the problem of event-triggered adaptive fault-tolerant fuzzy output feedback consensus tracking control for nonlinear fractional-order multiagent systems with actuator failures under a directed graph. Considering the fact that the actual system works near the equilibrium point most of the time, a novel dynamic event-triggering strategy with the reset mechanism is proposed, where the dynamic threshold can be actively adjusted according to the preset conditions, so that the resource utilization can be further reduced. Based on an improved event-based consensus error, the state estimator about the derivative of reference trajectory and the adaptive law about the information of graph are constructed, which makes distributed consensus tracking control achieved without obtaining global information. Then, by introducing two adaptive compensating terms to deal with actuator failures and event-triggered measurement errors, it is shown in the sense of fractional-order stability criterion that tracking errors can converge to a compact set even if the fault parameters and modes are completely unknown. Finally, the correctness of the presented method is verified by a simulation example.</p>","id":"09be7aa75e2b","score":9,"tags":["multi-agent systems","event-driven","graph theory","adaptive control","fault tolerance"],"keep":true,"rationale":"The paper addresses event-triggered adaptive control in multi-agent systems, which is highly relevant to the orchestration and adaptive aspects of the dissertation."} | |
| {"authors":["Wang, SM","Shu, Z","Chen, TW"],"type":"Journal Article","year":"2021","title":"Event-triggered attitude synchronization of multiple rigid-body systems","venue":"SYSTEMS & CONTROL LETTERS","volume":"149","doi":"10.1016/j.sysconle.2021.104879","wos_id":"WOS:000632569700006","abstract":"<p>In this paper, an attitude synchronization problem of multiple rigid-body systems is investigated by using an event-based approach. The leader and followers are described by unit quaternions. A nonlinear distributed observer with event-triggered observations is proposed to estimate the attitude and angular velocity of the leader without continuous information exchange. The triggering mechanism is intermittent and asynchronous; and a positive lower bound of inter-event triggering times is given to show that Zeno behavior can be excluded in the intermittent communication sequence for any agent. Based on the estimated attitude and angular velocity of the leader, a distributed controller is synthesized for each follower to achieve attitude synchronization with the leader via intermittent communication. Finally, an example is provided to illustrate the effectiveness of the theoretical results. (C) 2021 Elsevier B.V. All rights reserved.</p>","id":"c627b5d91735","score":8,"tags":["multi-agent","event-driven","synchronization","control systems"],"keep":true,"rationale":"The paper discusses an event-triggered approach for multi-agent systems, which aligns with the adaptive event-driven aspect of the dissertation."} | |
| {"authors":["Wanna, S","Parra, F","Valner, R","Kruusamaee, K","Pryor, M"],"type":"Journal Article","year":"2024","title":"Unlocking underrepresented use-cases for large language model-driven human-robot task planning","venue":"ADVANCED ROBOTICS","volume":"38","number":"18","pages":"1335-1348","doi":"10.1080/01691864.2024.2366974","wos_id":"WOS:001263051600001","abstract":"<p>Large language models (LLM) are now the de facto task planners for Embodied AI (EAI) systems. This shift can be attributed to LLMs' powerful, emergent properties which enable their adaptation to downstream tasks with minimal to no fine tuning via prompting. However, we find that LLM-driven task planning is not a solved problem. In this work we measure the extent to which these models can be adapted to complex and domain-specific task planning via few-shot prompting. Additionally, we contribute quantitative and qualitative analysis on prompt robustness. Lastly, to meet the challenges of adapting EAI systems to real-world, industrial domains, we adopt a human-in-the-loop approach to guarantee safe and interpretable task planning and execution. We successfully demonstrate co-located, human-robot teaming where an Augmented Reality (AR) headset mediates information exchanged between an EAI agent and human operator for a variety of inspection tasks. To our knowledge the use of an AR headset for multimodal grounding and the application of EAI to industrial tasks are novel contributions within Embodied AI research.</p>","id":"3b9c0ade7fed","score":8,"tags":["multi-agent","task planning","human-robot interaction","safety","embodied AI"],"keep":true,"rationale":"The paper discusses LLM-driven task planning in multi-agent systems with a focus on safety and human-robot interaction, which aligns well with the themes of adaptive orchestration and safety in multi-agent environments."} | |
| {"authors":["Willoughby, JR","Lamka, GF","Dunning, KH","Narine, L","Belsare, A","Sundaram, M"],"type":"Journal Article","year":"2025","title":"Using AI enhanced agent-based models to support management of wild populations","venue":"LANDSCAPE ECOLOGY","volume":"40","number":"7","doi":"10.1007/s10980-025-02149-2","wos_id":"WOS:001520899600014","abstract":"<p>ContextManaging wild populations in rapidly changing, human-dominated landscapes requires models that accommodate complex interactions among climate, land use, disease, and evolution. Agent-based models (ABMs) are well suited to this task but are often difficult to parameterize, calibrate, and interpret at management-relevant scales.ObjectivesWe discuss how artificial-intelligence (AI) techniques, including machine-learning regression, data-mining diagnostics, geospatial informatics, and large-language-model code aides, can streamline ABM parameter estimation and scenario testing, enhance extraction of decision-support metrics, and broaden the accessibility of ABMs for conservation planning.MethodsWe considered examples of AI use in ecology and evolution, including where AI was paired with ABMs, highlighting use cases such as calibration, rule discovery, data fusion, and code generation.ResultsWe show how supervised machine learning can supplement parameterization by learning relationships between empirical observations and model outputs. Data-mining methods may also be useful to identify parameters that drive most output variance. In addition, deep-learning remote-sensing products in ABMs allows landscape dynamics to be represented at ecologically relevant resolutions. Despite this, key obstacles, such as limited long-term ecological data, high computational demand, and the need for explainable-AI safeguards against biased predictions remain.ConclusionsExpanding the use of AI in ABMs will require interdisciplinary collaborations that pair ecologists with computer and geo-information scientists and explicit workflows for auditing AI decisions. However, leveraging AI enhanced ABMs will improve predictive modeling of species responses to environmental change, optimize conservation strategies, and develop more effective data-driven management.</p>","id":"f079e906a5ce","score":5,"tags":["agent-based models","AI","ecology","conservation","parameterization"],"keep":false,"rationale":"While the paper discusses agent-based models and AI, it focuses on ecological applications rather than multi-agent orchestration or event-driven systems relevant to the dissertation topic."} | |
| {"authors":["Wu, YH","Zhang, H","Wang, ZP"],"type":"Journal Article","year":"2023","title":"A robust hybrid triggered mechanism for consensus control of multi-agent systems under directed graphs","venue":"AUTOMATICA","volume":"154","doi":"10.1016/j.automatica.2023.111073","wos_id":"WOS:001001155800001","abstract":"<p>This paper intends to address the consensus problem of general linear multi-agent systems with unknown disturbances. To save resources and ensure robustness to unknown disturbances, a hybrid time-dynamic event triggered mechanism is designed, under which the time triggered mechanism is utilized to determine the explicit lower bound of inter-event time and a dynamic event-triggered mechanism is introduced to further avoid redundant triggering actions. Unlike most existing relevant literature, the stability analysis of systems and the procedure for the choice of parameters do not rely on conditions to which the existence of solutions is not proved. In addition, by the decomposition of weight-unbalanced directed graphs coupled with graph theory and matrix theory, the obtained results could be extended from strong strongly connected graphs to more general directed graphs only containing a spanning tree. Then, the corresponding robust asynchronous variable period sampling control strategy is also derived to avoid the continuous detection of triggering conditions. Finally, the effectiveness of the proposed algorithm is verified on the spacecraft formation flying problem.(c) 2023 Elsevier Ltd. All rights reserved.</p>","id":"948c375913c4","score":9,"tags":["multi-agent systems","event-driven","graph theory","consensus control","robustness"],"keep":true,"rationale":"The paper addresses consensus control in multi-agent systems using a hybrid event-triggered mechanism, which is highly relevant to the orchestration and safety aspects of the dissertation."} | |
| {"authors":["Xie, X","Sheng, T"],"type":"Journal Article","year":"2025","title":"Self-Triggered 6-DOF Formation Control for Multispacecraft Systems With Restricted Communication and Computation Resources","venue":"IEEE TRANSACTIONS ON AEROSPACE AND ELECTRONIC SYSTEMS","volume":"61","number":"2","pages":"4168-4177","doi":"10.1109/TAES.2024.3501233","wos_id":"WOS:001464949400006","abstract":"<p>The 6-DOF event-based formation control issue for multispacecraft systems considering unknown disturbances is investigated. First, a 6-DOF coupling dynamics is given, and a 6-DOF dynamic event-driven formation controller is designed to achieve stability and avoid continuous data transmission. The formation spacecraft updates its controller and transmits its states at the trigger instants. Further, a self-triggered function using triggering states to calculate the next triggering moment is designed to save computing resources. It should be noted that the logarithmic function does not need to be used in the computation of the next trigger moment, significantly improving resource conservation's effectiveness. Furthermore, the developed 6-DOF event-driven controllers make the system bounded stable and Zeno-free. Finally, simulations present that the self-triggered mechanism avoids the continuous computation of the trigger function while reducing the communication resource-saving efficiency by only 7% compared with the dynamic trigger mechanism.</p>","id":"0e6f2aea0256","score":8,"tags":["multi-agent","event-driven","formation control","resource management"],"keep":true,"rationale":"The paper discusses event-driven control mechanisms for multi-agent systems, which aligns well with the dissertation's focus on adaptive orchestration and resource management."} | |
| {"authors":["Xie, X","Sheng, T","He, L"],"type":"Journal Article","year":"2021","title":"Distributed Attitude Synchronization for Spacecraft Formation Flying via Event-Triggered Control","venue":"APPLIED SCIENCES-BASEL","volume":"11","number":"14","doi":"10.3390/app11146299","wos_id":"WOS:000676007500001","abstract":"<p>The distributed attitude synchronization control problem for spacecraft formation flying subject to limited energy and computational resources is addressed based on event-triggered mechanism. Firstly, a distributed event-driven controller is designed to achieve attitude coordination with the limitation of energy and computing resources. Under the proposed control strategy, the controller is only updated at the event triggering instants, which effectively reduces the update frequency. Subsequently, an event-triggered strategy is developed to further decrease energy consumption and the amount of computation. The proposed event-triggered function only requires the latest state information about its neighbors, implying that the trigger threshold does not need to be calculated continuously. It is shown that the triggering interval between two successive events is strictly positive, showing that the control system has no Zeno phenomenon. Moreover, the update frequency of the proposed controller can be reduced by more than 90% compared to the update frequency of the corresponding time-driven controller with an update frequency of 10 Hz by choosing appropriate control parameters and the control system can still achieve high-precision convergence. Finally, the effectiveness of the constructed control scheme is verified by numerical simulations.</p>","id":"77d1f3064bf0","score":8,"tags":["multi-agent","event-driven","synchronization","control","spacecraft"],"keep":true,"rationale":"The paper discusses event-triggered control mechanisms in a multi-agent context, which aligns well with the themes of adaptive event-driven orchestration."} | |
| {"authors":["Xiong, FB","Han, QH","Zhang, CN"],"type":"Journal Article","year":"2025","title":"Design AI Agent for Auditing: Applying Large Language Models (LLMs) and Retrieval Augmented Generations (RAG) to Audit Workflows","venue":"JOURNAL OF EMERGING TECHNOLOGIES IN ACCOUNTING","doi":"10.2308/JETA-2024-041","wos_id":"WOS:001578606300001","abstract":"<p>Foundation large language models (LLMs) face limitations in specialized auditing domains, including accuracy issues, contextual memory constraints, and manual document management requirements. This proposal introduces an AI agent framework specifically designed for auditing workflows, integrating three core components: retrieval augmented generation (RAG) for seamless access to private knowledge bases, customizable workflows with intelligent query classification and multiagent coordination, and orchestrated prompts that embed standardized audit methodologies. The proposed framework reduces workflow disruptions and token consumption while maintaining accuracy. The proposal demonstrates the agent's workflow and its capabilities in document retrieval and analytical calculations. The evaluation plans to compare foundation LLM applications with customized AI agents using both baseline RAG and graph RAG approaches across auditing tasks, measuring accuracy against manually generated ground truth and efficiency through time and token consumption metrics.</p>","id":"54b901cd5bc6","score":7,"tags":["multi-agent","workflow","LLM","orchestration","RAG"],"keep":true,"rationale":"The paper discusses an AI agent framework for auditing workflows that involves multi-agent coordination and customizable workflows, which aligns with the themes of orchestration and agent-based systems."} | |
| {"authors":["Xu, JY","Ding, BW","Peng, HE","Miao, W"],"type":"Journal Article","year":"2025","title":"Application of multi-agent reinforcement learning system in optimizing higher education resource allocation","venue":"JOURNAL OF COMPUTATIONAL METHODS IN SCIENCES AND ENGINEERING","doi":"10.1177/14727978251385187","wos_id":"WOS:001594590500001","abstract":"<p>Higher education resource allocation is confronted with significant systemic challenges, such as difficulties in coordinating multiple stakeholders, highly complex organizational structures, and a lack of dynamic adaptability-issues that critically undermine both the efficiency and equity of resource utilization. To tackle these challenges, this study introduces a novel Graph-based Multi-Agent Proximal Policy Optimization (GMAPPO) framework, enhanced with a graph attention mechanism, to enable intelligent and collaborative decision-making in higher education resource management. Designed to reflect the unique characteristics of the higher education ecosystem, the model employs a structured graph representation to precisely capture the intricate relationships among key entities-including institutions, academic programs, and faculty members. By incorporating a multi-head graph attention mechanism, the system improves its sensitivity to critical nodes, such as under-resourced or emerging disciplines, thereby enabling more targeted and adaptive resource allocation strategies. The system employs a centralized training and distributed execution framework, integrating global state perception with local autonomous decision-making within an actor-critic architecture. It achieves efficient collaboration between agents through policy gradient optimization. It innovatively applies a progressive course learning mechanism, manages resource conflicts and demand fluctuations in stages, and utilizes a phased loss smoothing strategy to jointly optimize global value goals and policy consistency, effectively balancing multiple educational goals such as efficiency and fairness. Experimental verification shows that compared to mainstream multi-agent algorithms such as MADDPG (Multi-Agent Deep Deterministic Policy Gradient), QMIX (Q-Mixing networks), and MAPPO (Multi-Agent Proximal Policy Optimization), the GMAPPO system achieves course and classroom resource utilization rates of 73.20% and 95.07%, respectively, in higher education scenarios. Its policy compliance rate reaches 98.5%. In a highly dynamic scenario with a 60% increase in new courses, its response time is only 3.2 s, highlighting its excellent real-time adaptability. Under normal load, its cost-effectiveness is 125.6 yuan per hour, and its average attention entropy is as low as 0.21, demonstrating its efficiency and decision focus in complex resource environments. Through structured feature extraction and dynamic collaborative optimization mechanisms, GMAPPO provides an efficient, fair, and robust intelligent system solution for higher education resource allocation.</p>","id":"0553625b8223","score":8,"tags":["multi-agent","graph-based","resource allocation","adaptive","optimization"],"keep":true,"rationale":"The paper discusses a multi-agent system with a graph-based approach for resource allocation, which aligns well with the themes of adaptive orchestration and decision-making in complex environments."} | |
| {"authors":["Xu, MY","Hao, F"],"type":"Journal Article","year":"2023","title":"Event-driven fully distributed optimal coordinated control for Euler-Lagrange multi-agent systems with connectivity preservation","venue":"JOURNAL OF THE FRANKLIN INSTITUTE","volume":"360","number":"13","pages":"10100-10126","doi":"10.1016/j.jfranklin.2023.07.031","wos_id":"WOS:001052832500001","abstract":"<p>This paper studies the distributed optimal coordinated control problem for Euler-Lagrange multiagent systems with connectivity preservation. The aim is to force agents to achieve the optimal solution minimizing the sum of the local objective functions while guaranteeing the connectivity of the communication graph. For practical purposes, the gradient vector of the local objective function is allowed to use only at the real-time generalized position instead of at the auxiliary system state. To make the control parameters independent of the global information and guarantee the fully distributed manner of controller, the adaptive control is introduced to update the coupling weights of the relative states among neighbors. Moreover, to reduce the resource for control updates, the event-driven communication is employed for the updates of both the relative states and the gradient of the connectivity-preserving potential function. Based on the Lyapunov analysis framework, it is proved that agents can converge to the optimal solution with connectivity preservation and Zeno behavior is excluded for the two event-triggering conditions. Finally, the effectiveness of the proposed method is verified by a numerical simulation example. & COPY; 2023 The Franklin Institute. Published by Elsevier Inc. All rights reserved.</p>","id":"f133a1fedaef","score":9,"tags":["multi-agent systems","event-driven","graph theory","adaptive control","connectivity preservation"],"keep":true,"rationale":"The paper focuses on event-driven control in multi-agent systems, emphasizing connectivity and optimal coordination, which aligns closely with the dissertation's themes."} | |
| {"authors":["Xu, T","Yang, T","Duan, ZS","Feng, G","Chen, GR"],"type":"Journal Article","year":"2023","title":"Distributed Coordination of Networked Manipulators: A Two-Layer Control Scheme","venue":"IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY","volume":"31","number":"6","pages":"2660-2672","doi":"10.1109/TCST.2023.3277595","wos_id":"WOS:001006572200001","abstract":"<p>This article studies the coordination control problem of networked manipulators, where manipulators are actuated to cooperatively transport an unknown object along a prescribed trajectory. It is assumed that the accurate position and velocity information of the trajectory is known only to a subset of manipulators. A two-layer control scheme consisting of an estimation layer and a control layer is developed. At the estimation layer, prescribed-time distributed estimators are designed to estimate the position and velocity of the trajectory. The estimation errors are shown to converge to zero in pre-specified time. At the control layer, an event-driven distributed controller with a dynamic triggering mechanism is proposed, where nonuniform time-varying communication delays are considered. A distinctive feature of the proposed controller is that both control updates and communication occur at certain event instants, and thus control update burden and communication costs are reduced. Moreover, two adaptive time-varying control gains are designed so that the controller can be implemented without requiring any inequality condition or global information. Finally, a simulation platform based on MATLAB Robotics Toolbox is designed and implemented to verify the effectiveness and superiority of the proposed two-layer control scheme.</p>","id":"bfa0aab357bb","score":8,"tags":["multi-agent","event-driven","control systems","coordination","adaptive"],"keep":true,"rationale":"The paper discusses a two-layer control scheme for networked manipulators with an event-driven approach, which aligns well with the themes of adaptive orchestration and multi-agent systems."} | |
| {"authors":["Yang, YR","Wang, PF","Liu, XM","Luo, WY","Yang, LB"],"type":"Journal Article","year":"2025","title":"A Multi-Agent and GraphRAG-Based Framework for Operation and Management Decision-Making in Hydraulic Projects","venue":"WATER RESOURCES MANAGEMENT","doi":"10.1007/s11269-025-04312-5","wos_id":"WOS:001530201800001","abstract":"<p>Inspection of hydraulic projects is a core management task to ensure the safe and stable operation of major infrastructure. The current traditional inspection model has significant shortcomings, such as high reliance on manual labor, low accuracy in identifying potential risks, and insufficient dynamic decision-making capabilities. This paper innovatively constructs a multi-agent collaborative intelligent decision-making framework that integrates large language models (LLMs) and graph retrieval-augmented generation (Graph RAG) technologies. Through a modular architecture encompassing perception, cognition, and decision, the framework achieves full-process automated inspection. The study employs multi-source inspection data from the past three years to construct a multimodal dataset. Domain-adaptive fine-tuning enhances the F1 scores of the multimodal large model in equipment recognition and defect detection by 7.2% and 6.9%, respectively. Furthermore, a dynamic knowledge graph system based on Graph RAG is established. Knowledge injection techniques compensate for gaps in domain-specific knowledge, while entity-relation reasoning mechanisms effectively mitigate model hallucination phenomena. Experimental results demonstrate that the hydraulic engineering inspection reports generated by this method, evaluated by both experts and operations personnel, accurately reflect professional knowledge and technical depth in the field of hydraulic engineering maintenance. This research provides a new technical paradigm with strong explainability and high reliability for the intelligent operation and maintenance of hydraulic engineering infrastructure, offering significant engineering application value to promote digital transformation within the industry.</p>","id":"02909d51bc35","score":8,"tags":["multi-agent","graph","decision-making","infrastructure","automation"],"keep":true,"rationale":"The paper discusses a multi-agent framework that integrates graph technologies for decision-making in infrastructure management, aligning well with the themes of adaptive orchestration and safety in multi-agent systems."} | |
| {"authors":["Yildirim, M","Dagda, B","Asodia, V","Fallah, S"],"type":"Journal Article","year":"2025","title":"HighwayLLM: Decision-making and navigation in highway driving with RL-informed language model","venue":"ROBOTICS AND AUTONOMOUS SYSTEMS","volume":"193","doi":"10.1016/j.robot.2025.105114","wos_id":"WOS:001524074400001","abstract":"<p>Autonomous driving is a complex task which requires advanced decision making and control algorithms. Understanding the rationale behind the autonomous vehicles' decision is crucial to ensure their safe and effective operation on highway driving. This study presents a novel approach, HighwayLLM, which harnesses the reasoning capabilities of large language models (LLMs) to predict the future waypoints for ego-vehicle's navigation. Our approach also utilizes a pre-trained Reinforcement Learning (RL) model to serve as a high-level planner, making decisions on appropriate meta-level actions. The HighwayLLM combines the output from RL model and the current state information to make safe, collision-free, and explainable predictions for next states, thereby constructing a trajectory for the ego-vehicle. Subsequently, a PID-based controller guides the vehicle to the waypoints predicted by the LLM agent. This integration of LLM with RL and PID enhances decision-making process, provides interpretability for highway autonomous driving and reduces the number of collisions compared to the baseline method.</p>","id":"56eb50f4154a","score":6,"tags":["autonomous driving","decision-making","reinforcement learning","safety"],"keep":true,"rationale":"While focused on autonomous driving, it discusses decision-making and safety, which are relevant to multi-agent orchestration and runtime verification."} | |
| {"authors":["Yu, H","Zeng, ZF","Li, K"],"type":"Journal Article","year":"2023","title":"Discrete-time event-based coordinated formation control of multiple AUVs with time-varying delay and alterable topology","venue":"PROCEEDINGS OF THE INSTITUTION OF MECHANICAL ENGINEERS PART I-JOURNAL OF SYSTEMS AND CONTROL ENGINEERING","volume":"237","number":"5","pages":"904-922","doi":"10.1177/09596518221137940","wos_id":"WOS:000898700400001","abstract":"<p>This article considers the issue of event-based coordinated formation control for multiple autonomous underwater vehicles subject to time-varying communication delay and alterable topology under discretized time. First, to derive the dynamics model in the discrete-time domain, the state feedback technique and the forward difference approach are employed for the continuous-time nonlinear model of autonomous underwater vehicle. For each follower, the event-triggering function and the coordinated controller are designed, using only discrete-time state information of the leader or other followers. Under this control strategy, the controller will not be updated at each discrete-time instant. Then, the coordinated control problem is turned into the asymptotic stability problem of the multi-autonomous underwater vehicle system. According to the stability analysis, sufficient conditions are presented for successfully completing the formation assignment without and with time-varying delay, respectively. Finally, by performing some numerical simulation experiments, the efficacity of our proposed event-based controller and the correctness of the main theoretical results are demonstrated. The impact of different communication conditions on the control system is revealed by the comparison of several scenarios.</p>","id":"78f7fdb173e1","score":8,"tags":["multi-agent","event-driven","formation control","autonomous systems"],"keep":true,"rationale":"The paper discusses event-based control for multiple autonomous vehicles, which aligns with the themes of multi-agent orchestration and event-driven systems."} | |
| {"authors":["Ze, KR","Wang, W","Liu, KX","Lu, JH"],"type":"Journal Article","year":"2024","title":"Time-Varying Formation Planning and Distributed Control for Multiple UAVs in Clutter Environment","venue":"IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS","volume":"71","number":"9","pages":"11305-11315","doi":"10.1109/TIE.2023.3335448","wos_id":"WOS:001125555500001","abstract":"<p>In this article, we present a novel method to achieve optimization based obstacle avoidance and distributed regular polygon time-varying formation control for multiple unmanned aerial vehicle systems (UAVs) in clutter environment. Under a leader-following structure, directed communication graph is considered. The formation size is time-varying and the trajectory of the leader UAV is planned in real time. Different from most of the existing trajectory planning algorithms for single UAV, an optimization-based safe trajectory and formation size online planning algorithm are proposed. To deal with information loss caused by directed communication topology, distributed smooth adaptive filters are designed for each UAV to asymptotically estimate the safe trajectory and formation size. Besides, a geometric tracking controller is adopted to track the desired trajectory for each UAV. Experimental results are provided to demonstrate the effectiveness of the proposed method.</p>","id":"24b03073cf79","score":8,"tags":["multi-agent","UAV","formation control","distributed control","safety"],"keep":true,"rationale":"The paper discusses distributed control and formation planning for multiple UAVs, which aligns with multi-agent orchestration and safety considerations."} | |
| {"authors":["Zhang, RX","Wang, BC","Zhang, JX","Bian, ZL","Feng, C","Ozbay, K"],"type":"Journal Article","year":"2025","title":"When language and vision meet road safety: Leveraging multimodal large language models for video-based traffic accident analysis","venue":"ACCIDENT ANALYSIS AND PREVENTION","volume":"219","doi":"10.1016/j.aap.2025.108077","wos_id":"WOS:001506047300001","abstract":"<p>The increasing availability of traffic videos functioning on a 24/7/365 time scale has the great potential of increasing the spatio-temporal coverage of traffic accidents, which will help improve traffic safety. However, analyzing footage from hundreds, if not thousands, of traffic cameras in a 24/7/365 working protocol still remains an extremely challenging task, as current vision-based approaches primarily focus on extracting raw information, such as vehicle trajectories or individual object detection, but require laborious post-processing to derive actionable insights. We propose SeeUnsafe, a new framework that integrates Multimodal Large Language Model (MLLM) agents to transform video-based traffic accident analysis from a traditional extraction-then-explanation workflow to a more interactive, conversational approach. This shist significantly enhances processing throughput by automating complex tasks like video classification and visual grounding, while improving adaptability by enabling seamless adjustments to diverse traffic scenarios and user-defined queries. Our framework employs a severity-based aggregation strategy to handle videos of various lengths and a novel multimodal prompt to generate structured responses for review and evaluation to enable fine-grained visual grounding. We introduce IMS (Information Matching Score), a new MLLM-based metric for aligning structured responses with ground truth. We conduct extensive experiments on the Toyota Woven Traffic Safety dataset, demonstrating that SeeUnsafe effectively performs accident-aware video classification and enables visual grounding by building upon off-the-shelf MLLMs. Our code will be made publicly available upon acceptance.</p>","id":"f503296e0c73","score":6,"tags":["multi-agent","LLM agents","workflow","safety","video analysis"],"keep":true,"rationale":"The paper discusses a framework that utilizes multimodal large language model agents for traffic accident analysis, which relates to multi-agent systems and adaptive workflows, though it focuses more on video analysis than orchestration."} | |
| {"authors":["Zhang, ZS","Yao, Y","Zhang, A","Tang, XR","Ma, XB","He, ZW","Wang, YM","Gerstein, M","Wang, R","Liu, GS","Zhao, H"],"type":"Journal Article","year":"2025","title":"Igniting Language Intelligence: The Hitchhiker's Guide from Chain-of-Thought Reasoning to Language Agents","venue":"ACM COMPUTING SURVEYS","volume":"57","number":"8","doi":"10.1145/3719341","wos_id":"WOS:001504536800002","abstract":"<p>Large language models (LLMs) have dramatically enhanced the field of language intelligence, as demonstrably evidenced by their formidable empirical performance across a spectrum of complex reasoning tasks. Additionally, theoretical proofs have illuminated their emergent reasoning capabilities, providing a compelling showcase of their advanced cognitive abilities in linguistic contexts. Critical to their remarkable efficacy in handling complex reasoning tasks, LLMs leverage the intriguing chain-of-thought (CoT) reasoning techniques, obliging them to formulate intermediate steps en route to deriving an answer. The CoT reasoning approach has not only exhibited proficiency in amplifying reasoning performance but also in enhancing interpretability, controllability, and flexibility. In light of these merits, recent research endeavors have extended CoT reasoning methodologies to nurture the development of autonomous language agents, which adeptly adhere to language instructions and execute actions within varied environments. This survey article orchestrates a thorough discourse, penetrating vital research dimensions, encompassing (i) the foundational mechanics of CoT techniques, with a focus on elucidating the circumstances and justification behind its efficacy; (ii) the paradigm shift in CoT; and (iii) the burgeoning of language agents fortified by CoT approaches. Prospective research avenues envelop explorations into generalization, efficiency, customization, scaling, and safety. A repository for the related papers is available at https://github.com/Zoeyyao27/CoT-Igniting-Agent.</p>","id":"ddc373ad7d95","score":7,"tags":["LLM","language agents","reasoning","autonomy","safety"],"keep":true,"rationale":"The paper discusses language agents and reasoning techniques that could inform the development of adaptive multi-agent systems, particularly in the context of orchestration and safety."} | |
| {"authors":["Zhu, HX","Su, HS"],"type":"Journal Article","year":"2025","title":"Event-Based Robust Adaptive Distributed Observer Under Directed Graphs","venue":"IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS","volume":"55","number":"8","pages":"5258-5269","doi":"10.1109/TSMC.2025.3565114","wos_id":"WOS:001484752300001","abstract":"<p>This article focuses on the event-based fully distributed state estimation problem under noisy environment and directed graphs. In a networked system, agents cooperatively estimate the target system state disturbed by process noise. However, due to the existence of measurement noise, each agent can only access partial and disturbed measurement output information. Meanwhile, given the limitations of resources in the networked system and the difficulty of acquiring global information, the event-based robust adaptive distributed observer is proposed. Specifically, by introducing robust adaptive coupling gains, a fully distributed design is achieved, eliminating the dependence on global information. Through the design of event-triggered mechanism, an event-based communication manner is implemented to reduce communication and energy resource consumption. Moreover, in view of the heterogeneity of the undetectable subspace and the differences in parameter design among agents, this article introduces a coordinate transformation and constructs a new Lyapunov function to further analyze the stability of the estimation error. Then, theoretical analysis shows that during cooperative estimation under directed graphs, the norm of estimation error can asymptotically converge to a compact set without Zeno behavior. Finally, a simulation example is provided to verify the effectiveness of the proposed event-based robust adaptive distributed observer.</p>","id":"9693488d0089","score":9,"tags":["multi-agent","event-driven","graph","adaptive","distributed systems"],"keep":true,"rationale":"The paper discusses event-based distributed state estimation in multi-agent systems under directed graphs, which aligns closely with the themes of adaptive orchestration and event-driven architectures."} | |
| {"authors":["Zhu, YA","Wen, GH","Yu, WW","Yu, XH"],"type":"Journal Article","year":"2021","title":"Nonsmooth Resource Allocation of Multiagent Systems With Disturbances: A Proximal Approach","venue":"IEEE TRANSACTIONS ON CONTROL OF NETWORK SYSTEMS","volume":"8","number":"3","pages":"1454-1464","doi":"10.1109/TCNS.2021.3068349","wos_id":"WOS:000696669000037","abstract":"<p>This article aims to solve the nonsmooth resource allocation problem in the presence of a global network resource constraint and local set constraints in the framework of multiagent system optimization. It is assumed that multiagent systems are subject to some external disturbances, and the control inputs of the agents satisfy Lispchitz continuity. These two distinguished features render the existing distributed optimization algorithms, especially the subgradient-based algorithms inapplicable due to the employment of discontinuity of subgradients. To solve such a challenging resource allocation problem, a new kind of continuous-time proximal algorithm is designed with the aid of convex optimization theory and the internal-model technique. The proximal algorithm is further augmented by introducing an event-based communication scheme such that the continuous-time communication among the agents is avoided successfully. The theoretical analysis shows that the multiagent systems under the proposed algorithms can converge to the optimal solution of the considered problem, while the external disturbances are rejected. Besides, the Zeno behavior can be excluded for the proximal algorithm with event-based communication. Finally, the numerical simulations are given to verify the established theoretical results.</p>","id":"fb04c86fa08a","score":8,"tags":["multi-agent systems","resource allocation","event-driven","optimization","disturbances"],"keep":true,"rationale":"The paper addresses resource allocation in multi-agent systems with an event-based communication scheme, which is relevant to adaptive orchestration and safety in agent runtime."} | |
| {"authors":["Zou, Y","Huang, Y","Xia, KW","Huang, BM","Dong, XF","Meng, ZY"],"type":"Journal Article","year":"2024","title":"Velocity-Free Distributed Optimization Algorithms for Second-Order Multiagent Systems","venue":"IEEE TRANSACTIONS ON CONTROL OF NETWORK SYSTEMS","volume":"11","number":"4","pages":"1911-1923","doi":"10.1109/TCNS.2024.3371550","wos_id":"WOS:001413492800011","abstract":"<p>This article concerns the distributed optimization problem of dynamical systems using partial state information. Such a problem is motivated by the fact that optimization missions are often subject to dynamical constraints and sensor malfunction. To cooperatively deal with the optimization problem for second-order multiagent systems in the absence of velocity information, we design two velocity-free distributed optimization algorithms over different communication topologies. First, for the case of the continuous communication, a fully velocity-free distributed optimization algorithm is designed by leveraging novel auxiliary dynamics, and each local cost function is just required to be convex. It is shown that all the agents are capable of achieving rendezvous on one of the optimal solutions of interest despite the absence of velocity information. Next, to relieve the communication burden, a modified velocity-free distributed optimization algorithm is designed by introducing an event-based communication mechanism, where all the local cost functions are required to be strongly convex. Particularly, a communication trigger condition is built such that the undesirable Zeno phenomenon is circumvented. Also, an adaptive gain is introduced to make the modified optimization algorithm fully distributed. Simulations are finally given to verify the optimization performance.</p>","id":"9139e7651767","score":8,"tags":["multi-agent","distributed optimization","event-driven","communication","safety"],"keep":true,"rationale":"The paper discusses distributed optimization in multi-agent systems, which is relevant for understanding orchestration and communication mechanisms in agent-based infrastructures."} | |
| {"authors":["Zou, Y","Wang, ZW","Huang, JS","Song, J","Xu, L"],"type":"Journal Article","year":"2025","title":"Multi-Agent Reinforcement Learning for Mobile Energy Resources Scheduling Amidst Typhoons","venue":"IEEE TRANSACTIONS ON INDUSTRY APPLICATIONS","volume":"61","number":"1","pages":"1683-1694","doi":"10.1109/TIA.2024.3463608","wos_id":"WOS:001410418500038","abstract":"<p>Substantial threats are posed by typhoon events to critical electrical power infrastructure that can result in human casualties and significant economic losses. Both traditional and renewable power generation systems can be negatively impacted, leading to widespread power outages that compromise public safety. In response, we introduce a novel spatial-topological multi-agent reinforcement learning (ST-MARL) method to optimize post-typhoon power system recovery by leveraging mobile energy resources (MERs). Compared to existing MARL methods, the proposed ST-MARL method utilizes convolutional neural network (CNN) and graph convolutional network (GCN) models to extract spatial-topological information from environmental data like typhoon meteorological conditions and power system states, which facilitates real-time decision-making for the routing and scheduling of MERs. Furthermore, this ST-MARL method achieves a two-stage decision-making process by allocating MERs' rewards with an Alternating Current Optimal Power Flow(AC-OPF) model. This ensures safety and feasibility of the decisions. Additionally, we employ the centralized training and distributed execution (CTDE) paradigm to address coordination challenges among MERs. These approaches collectively aim to enhance coordination among MERs, improve economic efficiency, and ensure the power supply of critical loads during post-typhoon power system recovery. Finally, in our case study of the Hong Kong power network, the results indicate that our ST-MARL method outperforms two main existing MARL methods, achieving an improvement of 5.13% and 6.77% on the reward, respectively.</p>","id":"e309ee04443e","score":8,"tags":["multi-agent","reinforcement learning","graph-based","safety","orchestration"],"keep":true,"rationale":"The paper discusses a multi-agent reinforcement learning approach that incorporates graph-based methods for decision-making in power system recovery, aligning well with the themes of adaptive orchestration and safety in multi-agent systems."} | |
| {"authors":["Zuzek, T","Vrabic, R","Zdesar, A","Skulj, G","Banfi, I","Bosnak, M","Zaletelj, V","Klancar, G"],"type":"Journal Article","year":"2024","title":"Simulation-Based Approach for Automatic Roadmap Design in Multi-AGV Systems","venue":"IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING","volume":"21","number":"4","pages":"6190-6201","doi":"10.1109/TASE.2023.3323099","wos_id":"WOS:001088286900001","abstract":"<p>This paper addresses the problem of establishing efficient intralogistic systems, focusing on the generation of roadmaps on a given layout and the coordination of multiple Automated Guided Vehicles (AGVs). A simulation-based approach for automatic roadmap design is proposed. An event-based simulator is developed that uses ant-colony inspired optimization to generate roadmaps tailored to the specific characteristics of a given intralogistic problem, i.e., the plant layout, fleet size, statistical description of tasks, dispatching algorithm, etc. The generated solutions are evaluated with a Multi-Agent Path Finding (MAPF) simulator that uses a Safe Interval Path Planning (SIPP) algorithm. By analysing the system throughput, the optimal fleet size for the system is proposed. The approach is validated through various examples and benchmarked against existing methods in the literature.</p>","id":"e6e6523959a1","score":8,"tags":["multi-agent","event-driven","graph","safety","orchestration"],"keep":true,"rationale":"The paper discusses multi-agent coordination and event-based simulation, which are relevant to the orchestration of agents in a graph-based infrastructure."} |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment