Optimal navigation for AGVs: A soft actor-critic-based reinforcement learning approach with composite auxiliary rewards.
Haisen GuoZhigang RenJialun LaiZongze WuShengli XiePublished in: Eng. Appl. Artif. Intell. (2023)
Keyphrases
- reinforcement learning
- actor critic
- optimal control
- approximate dynamic programming
- function approximation
- policy gradient
- average reward
- reinforcement learning algorithms
- markov decision processes
- temporal difference
- dynamic programming
- control policy
- state space
- model free
- policy iteration
- control problems
- learning algorithm
- optimal policy
- reward function
- supervised learning
- optimal solution
- multi agent
- gradient method
- temporal difference learning
- machine learning
- infinite horizon
- markov decision process
- learning problems
- stochastic games
- state action
- natural actor critic