Multiple Tree for Partially Observable Monte-Carlo Tree Search

TL;DR

This paper introduces a Monte-Carlo tree search algorithm leveraging bandit methods to compute approximate Nash equilibria in partially observable games, demonstrated on phantom tic-tac-toe.

Contribution

It presents a novel algorithm that efficiently computes strong strategies for partially observable games using Monte-Carlo tree search and bandit techniques.

Findings

Effective computation of approximate Nash equilibria.

Strong strategies achieved in phantom tic-tac-toe.

Algorithm shows promising results in partial observability settings.

Abstract

We propose an algorithm for computing approximate Nash equilibria of partially observable games using Monte-Carlo tree search based on recent bandit methods. We obtain experimental results for the game of phantom tic-tac-toe, showing that strong strategies can be efficiently computed by our algorithm.