Abstract:
In recent years, idempotent methods (specifically, max-plus methods) have been developed for solution of nonlinear control problems. We extend the applicability of idempo...Show MoreMetadata
Abstract:
In recent years, idempotent methods (specifically, max-plus methods) have been developed for solution of nonlinear control problems. We extend the applicability of idempotent methods to deterministic dynamic games through usage of the min-max distributive property. However, this induces a very high curse-of-complexity. A representation of the space of max-plus hypo-convex functions as a min-max linear space is used to obtain a result which may be used to attenuate this complexity growth. We apply this approach in a game of deception, where one player is searching for certain objects, while the other player may employ deception to hinder that search. The problem is formulated as a dynamic game, where the state space is a max-plus probability simplex.
Published in: Proceedings of the 2011 American Control Conference
Date of Conference: 29 June 2011 - 01 July 2011
Date Added to IEEE Xplore: 18 August 2011
ISBN Information: