The enhanced evolutionary tabu search and its application to the quadratic assignment problem

We describe the Enhanced Evolutionary Tabu Search (EE-TS) local search technique. The EE-TS metaheuristic technique combines Reactive Tabu Search with evolutionary computing elements proven to work well in multimodal search spaces. An initial set of solutions is generated using a stochastic heuristic operator based on Restricted Candidate List. Reactive Tabu Search is augmented with selection and recombination operators that preserve common traits between solutions while maintaining a diverse set of good solutions. EE-TS performance is applied to the Quadratic Assignment Problem using problem instances from the QAPLIB. The results show that EE-TS compares favorably against other known techniques. In most cases, EE-TS was able to find the known optimal solutions in fewer iterations. We conclude by describing the main benefits and limitations of EE-TS. Categories and Subject Descriptors I.2.8 [Artificial Intelligence]: Problem Solving, Control Methods, and Search – heuristics.

• Type: Conference
• Year: 2005
• Where: GECCO
• Authors: John F. McLoughlin III, Walter Cedeño
• Keywords: Evolutionary Tabu Search, GECCO 2005, Reactive Tabu Search, Tabu Search