Square-root Regret Bounds For Continuous-time Episodic Markov Decision Processes

Abstract

We study reinforcement learning for continuous-time Markov decision processes (MDPs) in the finite-horizon episodic setting. In contrast to discrete-time MDPs, the inter-transition times of a continuous-time MDP are exponentially distributed with rate parameters depending on the state--action pair at each transition. We present a learning algorithm based on the methods of value iteration and upper confidence bound. We derive an upper bound on the worst-case expected regret for the proposed algorithm, and establish a worst-case lower bound, both bounds are of the order of square-root on the number of episodes. Finally, we conduct simulation experiments to illustrate the performance of our algorithm.

Stats

• citations1
• S2 citations—
• github stars0
• HF likes0
• heat score2.26
• arxiv keygao2022square

Related papers

• Logarithmic Regret Bounds For Continuous-time Average-reward Markov Decision Processes (2022)5.24
• Logarithmic Regret For Episodic Continuous-time Linear-quadratic Reinforcement Learning Over A Finite-time Horizon (2020)7.81
• Online Reinforcement Learning In Markov Decision Process Using Linear Programming (2023)3.58
• Beyond Value-function Gaps: Improved Instance-dependent Regret Bounds For Episodic Reinforcement Learning (2021)0.00
• Logarithmic Regret Of Exploration In Average Reward Markov Decision Processes (2025)0.00
• Dynamic Regret Of Online Markov Decision Processes (2022)0.00
• Regret Analysis In Deterministic Reinforcement Learning (2021)0.00
• Efficient Learning In Non-stationary Linear Markov Decision Processes (2020)6.77