Exploring the Pareto front of multi-objective COVID-19 mitigation policies using reinforcement learning.
Mathieu ReymondConor F. HayesLander WillemRoxana RadulescuSteven AbramsDiederik M. RoijersEnda HowleyPatrick MannionNiel HensAnn NowéPieter LibinPublished in: CoRR (2022)
Keyphrases
- multi objective
- reinforcement learning
- optimal policy
- policy search
- evolutionary algorithm
- multi objective optimization
- multiobjective optimization
- markov decision process
- control policies
- optimization algorithm
- particle swarm optimization
- nsga ii
- reinforcement learning agents
- fitted q iteration
- reward function
- objective function
- pareto optimal
- markov decision processes
- hierarchical reinforcement learning
- control policy
- multi objective optimization problems
- function approximation
- state space
- reinforcement learning algorithms
- bi objective
- genetic algorithm
- multiple objectives
- partially observable markov decision processes
- markov decision problems
- pareto fronts
- learning algorithm
- dynamic programming
- multi agent
- model free
- policy gradient methods
- multi objective evolutionary algorithms
- long run
- conflicting objectives
- temporal difference
- action selection
- continuous state
- infinite horizon
- pareto dominance
- optimal control
- function approximators