Near-optimal Regret Bounds for Stochastic Shortest Path

Stochastic shortest path (SSP) is a well-known problem in planning and control, in which an agent has to reach a goal state in minimum total expected cost. In the learning formulation of the problem, the agent is unaware of the environment dynamics (i.e., the transition function) and has to repeatedly play for a given number of episodes while reasoning about the problem's optimal solution. Unlike other well-studied models in reinforcement learning (RL), the length of an episode is not predetermined (or bounded) and is influenced by the agent's actions. Recently, Tarbouriech et al. (2019) studied this problem in the context of regret minimization and provided an algorithm whose regret bound is inversely proportional to the square root of the minimum instantaneous cost. In this work we remove this dependence on the minimum cost---we give an algorithm that guarantees a regret bound of $\widetilde{O}(B_\star |S| \sqrt{|A| K})$, where $B_\star$ is an upper bound on the expected cost of the optimal policy, $S$ is the set of states, $A$ is the set of actions and $K$ is the number of episodes. We additionally show that any learning algorithm must have at least $\Omega(B_\star \sqrt{|S| |A| K})$ regret in the worst case.

[1]  Peter Auer,et al.  The Nonstochastic Multiarmed Bandit Problem , 2002, SIAM J. Comput..

[2]  Ambuj Tewari,et al.  REGAL: A Regularization based Algorithm for Reinforcement Learning in Weakly Communicating MDPs , 2009, UAI.

[3]  Rémi Munos,et al.  Minimax Regret Bounds for Reinforcement Learning , 2017, ICML.

[4]  András György,et al.  The adversarial stochastic shortest path problem with unknown transition probabilities , 2012, AISTATS.

[5]  Dimitri P. Bertsekas,et al.  Stochastic Shortest Path Problems Under Weak Conditions , 2013 .

[6]  E. Ordentlich,et al.  Inequalities for the L1 Deviation of the Empirical Distribution , 2003 .

[7]  Yishay Mansour,et al.  Online Convex Optimization in Adversarial Markov Decision Processes , 2019, ICML.

[8]  Yishay Mansour,et al.  Online Stochastic Shortest Path with Bandit Feedback and Unknown Transition Function , 2019, NeurIPS.

[9]  Emma Brunskill,et al.  Tighter Problem-Dependent Regret Bounds in Reinforcement Learning without Domain Knowledge using Value Function Bounds , 2019, ICML.

[10]  Peter Auer,et al.  Near-optimal Regret Bounds for Reinforcement Learning , 2008, J. Mach. Learn. Res..

[11]  Gábor Lugosi,et al.  Prediction, learning, and games , 2006 .

[12]  John N. Tsitsiklis,et al.  An Analysis of Stochastic Shortest Path Problems , 1991, Math. Oper. Res..

[13]  Alessandro Lazaric,et al.  No-Regret Exploration in Goal-Oriented Reinforcement Learning , 2019, ICML.

[14]  Benjamin Van Roy,et al.  Generalization and Exploration via Randomized Value Functions , 2014, ICML.

[15]  Tor Lattimore,et al.  Unifying PAC and Regret: Uniform PAC Bounds for Episodic Reinforcement Learning , 2017, NIPS.

[16]  Gergely Neu,et al.  Online learning in episodic Markovian decision processes by relative entropy policy search , 2013, NIPS.