A Multi-Region Differential Evolution approach for continuous optimization problems

This paper presents a Multi-Region Differential Evolution (MRDE) algorithm as an extension of a classical version of differential evolution (DE) (i.e., as an extension of DE/rand-to-best/1/exp). MRDE is designed to simultaneously search on different and evenly distributed sub-regions on the whole search space. The number and extent of the search regions change during the execution of the algorithm, in such a way that, at the final stage of the evolutionary process, only one region remains (i.e., the whole search space). Our proposed MRDE is compared with respect to the classical DE algorithm on a set of well-known benchmark problems. The results achieved show enough evidence of the benefits of distributing the population of vectors when dealing with large-scale optimization problems.