Tabu Search
摘要
This chapter contains a discussion of the “Tabu Search”, a metaheuristic that explores the search space with a strategy avoiding to revisit solutions already seen, or with properties already known, at least within a predefined memory range. Pseudo-code and concepts are presented in a detailed way, trying to highlight the justification of the approach. The Tabu Search is a perfect illustration of the general mechanisms of metaheuristics. The chapter introduces the concept of “convergence”, which relates to the capability of a metaheuristics to find a global optimum. The case of Quadratic Optimization Problems (QAP) is used as an illustration of the Tabu Search method.