Enforcing Hard Constraints with Soft Barriers: Safe Reinforcement Learning in Unknown Stochastic Environments.
Yixuan WangSimon Sinong ZhanRuochen JiaoZhilu WangWanxin JinZhuoran YangZhaoran WangChao HuangQi ZhuPublished in: CoRR (2022)
Keyphrases
- hard constraints
- reinforcement learning
- soft constraints
- direct policy search
- cost function
- efficient computation
- constraint satisfaction problems
- graph cuts
- constraint satisfaction
- constraint violations
- multi objective evolutionary
- search space
- multi criteria
- constraint propagation
- dynamic programming
- state space
- machine learning
- evolutionary algorithm
- penalty function
- similarity search
- np hard
- genetic algorithm