Cuckoo Search Algorithm With Neighborhood Attraction for Numerical Optimization

Cuckoo search (CS) algorithm has been proved to be an effective method in solving numerical optimization problems. Nevertheless, with regard to Levy flight, each individual is attracted by the best solution found so far in the entire population, which may lead to premature convergence. Motivated by this observation, a new cuckoo search with neighborhood attraction (NACS) is proposed. In NACS, the neighborhood attraction scheme based on ring topology is firstly designed, where the best solution in a predefined neighborhood is employed to guide individual evolution. Then, to further enhance the exploration ability, the neighborhood attraction scheme and Levy flight are combined to generate potential candidate solutions. Moreover, the step size is adaptively regulated according to the degree of individual evolution. To validate the effectiveness of the presented algorithm, 25 extensively used benchmark test problems with different dimensions are employed. Experimental results reveal that the presented method is a competitive optimizer compared with other algorithms.