drl-supply-chain-hilo-dqn-rllib.py

import ray
import ray.rllib.agents.dqn as dqn

def train_dqn():
  config = dqn.DEFAULT_CONFIG.copy()
  config["log_level"] = "WARN"
  config["train_batch_size"] = 256
  config["buffer_size"] = 10000
  config["hiddens"] = [128, 128, 128]
  trainer = dqn.DQNTrainer(config=config, env=HiLoPricingEnv)
  for i in range(50):
    result = trainer.train()
    print(pretty_print(result))

train_dqn()