Niching Imperialistic Competitive Algorithm

Evolutionary computing techniques are among powerful tools for solving optimization problems. Imperialistic competitive algorithm is an evolutionary optimization algorithmthat has been shown to be successful for locating global optima. However in real world problems, it is critical to locate multiplesatisfactory solutions. This paper describes an optimization algorithm based on Imperialistic competitive algorithm for solving multimodal problems. The proposed method avoids using any prior knowledge about landscape of problem. It is capable of locating and maintaining found local and global optima duringthe run time. In order to avoid using any parameter for determining the size of niches within the landscape, algorithmadaptively adjusts its parameter according to the current population’s information. The proposed algorithm is tested on nine commonly used multimodal problems. Experimental results suggest that the proposed method is able to provide a superior performance or the same performance in some test functions over other existing multimodal algorithms

Keywords:

evolutionary computation , niching , multi-modal evolutionary optimization , algorithm , Imperialist Competitive Algorithm