Tabu Search Background

Abstract

The abundance of difficult optimization problems encountered in practical settings such as telecommunications, logistics, financial planning, transportation, and production has motivated the development of powerful optimization techniques. These techniques are usually the result of adapting ideas from a variety of research areas. The hope is to develop procedures that are efficient and are able to handle the complexity oftoday's optimization problems. The ideas that motivate the particular structure of each methodology sometimes come from unexpected connections. Network flow programming, for example, shares a heritage with models exploiting ideas from electricity and hydraulics. Simulated annealing (SA) is based on a physical process in metallurgy. Genetic algorithms (GAs) seek to imitate the biological phenomenon of evolutionary reproduction, while ant systems simulate a colony of ants that cooperate in a common problem solving activity. The philosophy of tabu search (TS) is to derive and exploit a collection of principles of intelligent problem solving. In this sense, it can be said that tabu search is based on selected concepts that unite the fields of artificial intelligence and optimization. The basic form of TS is founded on ideas proposed by Fred Glover. (See Section 1.8 for a brief sketch of historical origins.) The method is based on procedures designed to cross boundaries of feasibility or local optimality, which were usually treated as barriers. Early example of these procedures (derived from surrogate constraint methods and cutting plane approaches) systematically imposed and released constraints to permit exploration of otherwise forbidden regions. Seminal related ideas were also developed by Pierre Hansen in what he labeled a steepest ascent / mildest descent method (see Figure 1.1). In addition, as will become clear in this F. Glover et al.,