stefanopalmieri

import numpy as np
import gym
from gym import wrappers

env = gym.make('CartPole-v1')
env = wrappers.Monitor(env, '/tmp/cartpole-experiment-v1', force=True)

for i_episode in range(100):
    observation = env.reset()
    while True:

stefanopalmieri

import numpy as np
import math
from population import Population
from cartographer import Cartographer
import gym
from gym import wrappers

# NEAT experiment using the CAMPN encoding
# In this experiment node types used are Input, Output, Hadamard and Kronecker

stefanopalmieri

# A neural network is trained using NEAT
# The idea is from the paper: "Evolving Neural Networks through Augmenting Topologies"
# This gist is using MultiNEAT (http://multineat.com/)

import logging
import numpy as np
import pickle

import gym

stefanopalmieri

# A neural network is trained using ES-HyperNEAT
# The idea is from the paper: "Evolving Neural Networks through Augmenting Topologies"
# This gist is using MultiNEAT (http://multineat.com/)

import logging
import numpy as np
import pickle

import gym

stefanopalmieri

# A neural network is trained using ES-HyperNEAT
# The idea is from the paper: "Evolving Neural Networks through Augmenting Topologies"
# This gist is using MultiNEAT (http://multineat.com/)

import logging
import numpy as np
import pickle

import gym

stefanopalmieri

# Using ES-HyperNEAT to try to solve the Bipedal walker.
# This attempt was not successful. Adjustment of hyperparameters is likely needed.

# A neural network is trained using NeuroEvolution of Augmenting Topologies
# The idea is from the paper: "Evolving Neural Networks through Augmenting Topologies"
# This gist is using MultiNEAT (http://multineat.com/)

import logging
import numpy as np
import pickle

stefanopalmieri

# Solution organism was found in 78th generation
# Total of 200 * 5 * 78 = 78000 trials were used for training


from __future__ import print_function

import gym
import numpy as np
import itertools
import os

stefanopalmieri

#--- parameters for the pole experiment v1 ---#

# The `Types` section specifies which classes should be used for various
# tasks in the NEAT algorithm.  If you use a non-default class here, you
# must register it with your Config instance before loading the config file.
[Types]
stagnation_type      = DefaultStagnation
reproduction_type    = DefaultReproduction

[phenotype]

stefanopalmieri

#--- parameters for the pole experiment v1 ---#

# Two trials were held for each organism and the average fitness was taken.

# The `Types` section specifies which classes should be used for various
# tasks in the NEAT algorithm.  If you use a non-default class here, you
# must register it with your Config instance before loading the config file.
[Types]
stagnation_type      = DefaultStagnation
reproduction_type    = DefaultReproduction

stefanopalmieri

#--- parameters for the pole experiment ---#

# The `Types` section specifies which classes should be used for various
# tasks in the NEAT algorithm.  If you use a non-default class here, you
# must register it with your Config instance before loading the config file.
[Types]
stagnation_type      = DefaultStagnation
reproduction_type    = DefaultReproduction

[phenotype]