Adaptive Cuckoo Search Algorithm Based on Dynamic Scaling Factor and Historical Experience

Abstract

Cuckoo search (CS) algorithm is a simple and efficient bionic algorithm, it has advantages in solving the resources scheduling problem and optimize computer aided design. A novel adaptive CS algorithm, EDCS, combining historical experience and dynamic step size, is proposed to solve the problems of slow convergence and local optimization in the later stage of iteration. Through adaptive adjustment of Lévy flights step size, the algorithm has a large space for optimization in the early stage and thus improves the global search ability. In the later stage, the step size decreases with iteration and the local search ability of the algorithm is enhanced. Aiming at the problem that the overstepping nest caused by the preferential random walks affects the convergence speed and algorithm accuracy, a memory strategy is introduced to make the algorithm in full use of the historical iterative experiences, which improves the stability. The experimental results show that the performance of EDCS algorithm is significantly better than that of the original algorithm.