Reward penalties on augmented states for solving richly constrained RL effectively

• Main Authors: HAO, Jiang, MAI, Tien, VARAKANTHAN, Pradeep, HOANG, Minh Huy
• Format: text
• Language: English
• Published: Institutional Knowledge at Singapore Management University 2024
• Subjects: Safe reinforcement learning; Reward penalties; Constraint optimization; Reinforcement learning; Markov models (MDPs; POMDPs); Stochastic optimization; Artificial Intelligence and Robotics
• Online Access: https://ink.library.smu.edu.sg/sis_research/9685; https://ink.library.smu.edu.sg/context/sis_research/article/10685/viewcontent/29962_Article_Text_34016_1_2_20240324.pdf
• Institution: Singapore Management University

Constrained Reinforcement Learning employs trajectory-based cost constraints (such as expected cost, Value at Risk, or Conditional VaR cost) to compute safe policies. The challenge lies in handling these constraints effectively while optimizing expected reward. Existing methods convert such trajecto...