Algorithms, hardness and graph products on a pursuit-evasion game

TL;DR

This paper studies a pursuit-evasion game on graphs, providing algorithms for certain cases and proving NP-hardness in others, advancing understanding of the game's computational complexity and strategies.

Contribution

It introduces a polynomial-time algorithm for fixed guard numbers and parameters, and proves NP-hardness for bipartite graphs with bounded diameter.

Findings

Polynomial-time algorithm for fixed number of guards with parameters s≥2, d≥0

NP-hardness of the game on bipartite graphs with bounded diameter

Fixed parameter tractability related to P4-fewness of the graph

Abstract

In the $(s,d)$-spy game over a graph, introduced by Cohen et al. in 2016, one spy and $k$ guards occupy vertices of a graph and, at each turn, each guard may move along one edge and the spy may move along at most $s$ edges. The guards win if, after a finite number of turns, they ensure that the spy always remains at distance at most $d$ from at least one guard. The guard number is the minimum number of guards such that the guards have a winning strategy. In this paper, we investigate the spy game variant in which the guards are placed first, before the spy. We obtain a polynomial time algorithm for every speed $s\geq 2$ and distance $d\geq 0$ when the number of guards is a constant, which leads to a fixed parameter tractable algorithm on the $P_4$-fewness of the graph. We also prove that the spy game is NP-hard even in bipartite graphs with bounded diameter, for every speed $s\geq 2$ and distance $d\geq 0$.