An intelligent scatter with estimation of distribution for Tabu search

Abstract

This paper proposes an efficient optimization algorithm based on tabu search and estimation of distribution algorithms. The proposed algorithm estimates characteristics of distribution of solutions after performing tabu searches based on a marginal product model to acquire linkage information. Once correct linkage information is obtained, we can perform crossovers effectively without disrupting building blocks using the information. The proposed algorithm is expected to adapt to both global and local characteristics of solution space. Through empirical studies, we show the effectiveness of our approach. © Springer-Verlag Berlin Heidelberg 2007.