Control scaling factor of cuckoo search algorithm using learning automata

Abstract

In this study, we seek an optimal scaling factor of cuckoo search algorithm by using learning automata. In the presented method, the same learning automaton is built for each individual, and a set of actions of each learning automaton are set to several constant scaling factors. Moreover, the linear reward-penalty learning algorithm is used in learning automaton to select the optimal scaling factor of each individual. Extensive experiments on 20 benchmark functions demonstrate better effectiveness and efficiency of controlling scaling factor of cuckoo search by using learning automata.