Enhanced differential evolution hybrid scatter search for discrete optimization

Abstract

A hybrid approach of the enhanced differential evolution (EDE) and scatter search (SS), termed HEDE-SS, is presented in order to solve discrete domain optimization problems. This approach is envisioned in order to capture the randomization properties of EDE and the memory adaptation programming (MIAP) properties of SS. Two highly demanding problems of quadratic assignment problem (QAP) and traveling salesman problem (TSP) are optimized with this new heuristic approach. The hybrid obtains the optimal results for almost all of the QAP instances, compares very well for symmetric TSP by getting results around 98 per cent to the optimal.