Nearly Minimax Optimal Reinforcement Learning for Linear Mixture Markov Decision Processes

We study reinforcement learning (RL) with linear function approximation where the underlying transition probability kernel of the Markov decision process (MDP) is a linear mixture model (Jia et al., 2020; Ayoub et al., 2020; Zhou et al., 2020) and the learning agent has access to either an integration or a sampling oracle of the individual basis kernels. We propose a new Bernstein-type concentration inequality for self-normalized martingales for linear bandit problems with bounded noise. Based on the new inequality, we propose a new, computationally efficient algorithm with linear function approximation named for the aforementioned linear mixture MDPs in the episodic undiscounted setting. We show that attains an regret where is the dimension of feature mapping, is the length of the episode and is the number of interactions with the MDP. We also prove a matching lower bound for this setting, which shows that is minimax optimal up to logarithmic factors. In addition, we propose the algorithm for the same family of MDPs under discounting and show that it attains an regret, where is the discount factor. Our upper bound matches the lower bound proved by Zhou et al. (2020) up to logarithmic factors, suggesting that is nearly minimax optimal. To the best of our knowledge, these are the first computationally efficient, nearly minimax optimal algorithms for RL with linear function approximation.
View on arXiv