A Study on Hybrid Optimization Methods using Lévy Flight and Differential Evolution Mechanisms

Cuckoo search (CS) and differential evolution (DE) are multipoint search metaheuristics proposed for solving continuous optimization problems. Both methods exhibit relatively high performances compared to similar methods. In addition, they are easy to use because few parameters are preset by users. Moreover, they share the common feature of being easy to implement because they are simple algorithms. In contrast, CS is a global search-type method, and DE is a local search-type method. In this study, we propose a hybrid method that combines the generation update of DE with the Lévy flight of CS. We also conducted numerical experiments using a benchmark test suite, compared its performance with that of conventional CS and DE, and confirmed the effectiveness of the proposed method.