Searching for an evader in an unknown graph by an optimal number of searchers

Abstract

The graph search problem is the problem of searching a graph G for a mobile evader by mobile searchers. The edge search is an offline and centralized version, and es(G) denotes the number of searchers necessary and sufficient to edge search G. An online and distributed setting assumes a port numbering of G, a distinct homebase and a whiteboard in each node. Search algorithms typically respect the monotone and connected search strategy to protect the information on whiteboards; however, Ω(n/ log n es(G)) searchers are necessary even for trees, where n is the order of G. We investigate the problem under a new online and distributed setting: We assume that searchers can exchange information wherever they meet, instead of assuming a port numbering, a homebase and whiteboards. Under this setting, we propose a search algorithm for es(G) searchers, which is optimal.