The Explorer-Director Game on Graphs

TL;DR

This paper analyzes the Explorer-Director game on graphs, providing structural insights into optimal strategies, and offers complete solutions for trees and lattices, advancing understanding of this adversarial pursuit game.

Contribution

It reduces the analysis of the game to minimal closed vertex sets and solves the game on trees and lattices, including strategy restrictions.

Findings

Structural characterization of optimal strategies.

Complete solution for trees under restricted strategies.

Analysis of the game on lattice graphs.

Abstract

The Explorer-Director game, first introduced by Nedev and Muthukrishnan, can be described as a game where two players -- Explorer and Director -- determine the movement of a token on the vertices of a graph. At each time step, the Explorer specifies a distance that the token must move hoping to maximize the amount of vertices ultimately visited, and the Director adversarially chooses where to move token in an effort to minimize this number. Given a graph and a starting vertex, the number of vertices that are visited under optimal play is denoted by $f_d(G,v)$. In this paper, we first reduce the study of $f_d (G,v)$ to the determination of the minimal sets of vertices that are \textit{closed} in a certain combinatorial sense, thus providing a structural understanding of each player's optimal strategies. As an application, we address the problem on lattices and trees. In the case of trees, we also provide a complete solution even in the more restrictive setting where the strategy used by the Explorer is not allowed to depend on their opponent's responses. In addition to this paper, a supplementary companion note will be posted to arXiv providing additional results about the game in a variety of specific graph families.