Asynchronous Vector Iteration in Multi-objective Markov Decision Processes.
Ekaterina SedovaLawrence MandowJosé-Luis Pérez-de-la-CruzPublished in: CAEPIA (2021)
Keyphrases
- markov decision processes
- multi objective
- objective function
- evolutionary algorithm
- multi objective optimization
- reinforcement learning
- optimal policy
- state space
- finite state
- dynamic programming
- genetic algorithm
- policy iteration
- transition matrices
- particle swarm optimization
- multiple objectives
- decision theoretic planning
- nsga ii
- action sets
- planning under uncertainty
- model based reinforcement learning
- factored mdps
- partially observable
- finite horizon
- average reward
- average cost
- neural network
- reachability analysis
- decision processes
- reinforcement learning algorithms
- infinite horizon
- pareto optimal
- action space
- state and action spaces
- reward function
- semi markov decision processes
- linear programming
- least squares
- discounted reward