Modified cuckoo search algorithm and the prediction of flashover voltage of insulators

Abstract

A novel modified cuckoo search (MCS) algorithm is proposed to enhance the performance of cuckoo search. In this MCS algorithm, an enhancing exploration strategy is adopted to extend the search region. At the same time, the control parameters of step size and discovery probability are adjusted adaptively according to the improvement rate of the solutions toward the optimal value. To evaluate the effectiveness of the presented MCS algorithm, the simulation experiment is executed on 23 benchmark test functions. The results demonstrate that MCS algorithm performs better than the others reported in the literature. Moreover, a combination algorithm, which the MCS combined with BP neural network, is utilized to predict the flashover voltage on polluted insulators. The comparative study results further indicate that MCS is more efficient and effective, and the combination algorithm is suitable for the prediction of flashover voltage on insulators.