Differential Evolution with Clustering Cooperative Coevolution for High-Dimensional Problems

Abstract

Recently, evolutionary algorithms have been successful to solve many optimization problems. However, their performance will deteriorate when applied to complex high-dimensional problems. A clustering-cooperative coevolution scheme was introduced into DE algorithm to tackle the high-dimensional problems. In the scheme, the clustering method has been employed to decompose the problem, which works well with the cooperative coevolution. The proposed algorithm is evaluated by MPB and CEC09 benchmark functions with expanded dimension. The results are very promising, which show clearly that our proposed algorithm is effective for dynamic high-dimensional optimization problems.