Enhancing the cuckoo search with levy flight through population estimation

This paper proposed the use of population estimation in a new meta-heuristic called Cuckoo search (CS) algorithm to minimize the training error, achieve fast convergence rate and to avoid local minimum problem. The CS algorithm which imitates the cuckoo bird’s search behavior for finding the best nest has been applied independently to solve several engineering design optimization problems based on cuckoo bird’s behavior. The algorithm is tested on five benchmark functions such as Ackley function, Griewank function, Rastrigin function, Rosenbrock function and Schwefel function. The performance of the proposed algorithm was compared with Particle Swarm Optimization (PSO), Wolf Search Algorithm (WSA) and Artificial Bee Colony (ABC). The simulation results show that the CS with Levy flight out performs PSO, WSA and ABC, when the cuckoo population is varied