The Impact of Adversarial Knowledge on Adversarial Planning in Perimeter Patrol

DMCA

The Impact of Adversarial Knowledge on Adversarial Planning in Perimeter Patrol (2008)

Cached

Download Links

by Noa Agmon , Vladimir Sadov , Gal A. Kaminka , Sarit Kraus

Citations:

34 - 14 self

BibTeX

@MISC{Agmon08theimpact,
    author = {Noa Agmon and Vladimir Sadov and Gal A. Kaminka and Sarit Kraus},
    title = {The Impact of Adversarial Knowledge on Adversarial Planning in Perimeter Patrol },
    year = {2008}
}

OpenURL

Abstract

This paper considers the problem of multi-robot patrolling around a closed area, in the presence of an adversary trying to penetrate the area. Previous work on planning in similar adversarial environments addressed worst-case settings, in which the adversary has full knowledge of the defending robots. It was shown that non deterministic algorithms may be effectively used to maximize the chances of blocking such a full-knowledge opponent, and such algorithms guarantee a “lower bound” to the performance of the team. However, an open question remains as to the impact of the knowledge of the opponent on the performance of the robots. This paper explores this question in depth and provides theoretical results, supported by extensive experiments with 68 human subjects concerning the compatibility of algorithms to the extent of information possessed

Keyphrases

adversarial knowledge adversarial planning perimeter patrol full knowledge full-knowledge opponent worst-case setting algorithm guarantee open question previous work theoretical result closed area non deterministic algorithm similar adversarial environment extensive experiment human subject

DMCA