Multi-objective Reinforcement Learning in Factored MDPs with Graph Neural Networks.
Marc VincentAmal El Fallah SeghrouchniVincent CorrubleNarayan BernardinRami KassabFrédéric BarbarescoPublished in: AAMAS (2023)
Keyphrases
- factored mdps
- multi objective
- reinforcement learning
- markov decision processes
- neural network
- state space
- approximate dynamic programming
- policy iteration
- genetic algorithm
- multi objective optimization
- evolutionary algorithm
- reinforcement learning algorithms
- function approximators
- context specific
- algebraic decision diagrams
- markov decision problems
- model free
- function approximation
- dynamic programming
- particle swarm optimization
- multiple objectives
- transition model
- optimal policy
- objective function
- random walk
- finite state
- action space
- markov decision process
- planning under uncertainty
- linear programming
- pareto optimal
- supervised learning
- basis functions
- learning capabilities
- differential evolution
- multi agent
- nsga ii