Multi-objective Reinforcement Learning through Continuous Pareto Manifold Approximation.
Simone ParisiMatteo PirottaMarcello RestelliPublished in: J. Artif. Intell. Res. (2016)
Keyphrases
- multi objective
- reinforcement learning
- multi objective optimization
- evolutionary algorithm
- multiobjective optimization
- optimization algorithm
- continuous functions
- action space
- pareto optimal
- genetic algorithm
- multiple objectives
- function approximation
- objective function
- particle swarm optimization
- piecewise constant
- conflicting objectives
- state space
- closed form
- nsga ii
- model free
- multi objective optimization problems
- continuous state spaces
- manifold learning
- pareto optimal solutions
- learning algorithm
- high dimensional
- continuous domains
- continuous state and action spaces
- bi objective
- neural network
- multi objective evolutionary
- reinforcement learning algorithms
- temporal difference
- low dimensional
- differential evolution
- markov decision processes
- geodesic distance
- optimal control
- linear programming
- approximation algorithms
- real valued
- parameter space
- gaussian convolution
- optimal policy