Distributed Asynchronous Policy Iteration for Sequential Zero-Sum Games and Minimax Control

07/22/2021
by   Dimitri Bertsekas, et al.
0

We introduce a contractive abstract dynamic programming framework and related policy iteration algorithms, specifically designed for sequential zero-sum games and minimax problems with a general structure. Aside from greater generality, the advantage of our algorithms over alternatives is that they resolve some long-standing convergence difficulties of the "natural" policy iteration algorithm, which have been known since the Pollatschek and Avi-Itzhak method [PoA69] for finite-state Markov games. Mathematically, this "natural" algorithm is a form of Newton's method for solving Bellman's equation, but Newton's method, contrary to the case of single-player DP problems, is not globally convergent in the case of a minimax problem, because the Bellman operator may have components that are neither convex nor concave. Our algorithms address this difficulty by introducing alternating player choices, and by using a policy-dependent mapping with a uniform sup-norm contraction property, similar to earlier works by Bertsekas and Yu [BeY10], [BeY12], [YuB13]. Moreover, our algorithms allow a convergent and highly parallelizable implementation, which is based on state space partitioning, and distributed asynchronous policy evaluation and policy improvement operations within each set of the partition. Our framework is also suitable for the use of reinforcement learning methods based on aggregation, which may be useful for large-scale problem instances.

READ FULL TEXT

page 1

page 2

page 3

page 4

05/04/2020

Multiagent Value Iteration Algorithms in Dynamic Programming and Reinforcement Learning

We consider infinite horizon dynamic programming problems, where the con...
02/17/2021

Provably Efficient Policy Gradient Methods for Two-Player Zero-Sum Markov Games

Policy gradient methods are widely used in solving two-player zero-sum g...
01/19/2022

Anytime PSRO for Two-Player Zero-Sum Games

Policy space response oracles (PSRO) is a multi-agent reinforcement lear...
06/25/2020

Newton-type Methods for Minimax Optimization

Differential games, in particular two-player sequential games (a.k.a. mi...
12/12/2021

On the Heterogeneity of Independent Learning Dynamics in Zero-sum Stochastic Games

We analyze the convergence properties of the two-timescale fictitious pl...
10/06/2019

Biased Aggregation, Rollout, and Enhanced Policy Improvement for Reinforcement Learning

We propose a new aggregation framework for approximate dynamic programmi...