DEFINING TABU TENURE FOR THE QUADRATIC ASSIGNMENT PROBLEM

Abstract

Tabu search (TS) algorithms are among the most efficient heuristic techniques in combinatorial optimization. Within these algorithms, it is important that the proper policies for maintaining the tabu tenure (tabu list size) are applied. In this paper, we discuss the mechanisms of defining the tabu tenure for the famous combinatorial optimization problem − the quadratic assignment problem (QAP). Several variants of maintaining the tabu tenure are implemented. They are examined on the instances taken from the QAP instances library − QAPLIB. The results from the experiments demonstrate that the deterministic strategies of defining the tabu tenure are preferable to the corresponding random strategies.