Global optimisation using Pareto cuckoo search algorithm

Cuckoo search is one of nature-inspired algorithms successfully used for solving different optimisation problems. Cuckoo search has proved to be very effective than other nature-inspired algorithms, however, there is still room to improve it further by either by step sizes or probability of finding foreign egg. In this paper, we present an improved cuckoo search algorithm using a Pareto distribution instead of Levy distribution as per original cuckoo search. Five cuckoo search algorithms based on different distribution functions are developed and compared for performances, computational time and convergence rates. These cuckoo search algorithms performances were validated against ten standard test functions. It was found that the cuckoo search algorithm based on Pareto distribution outperformed other cuckoo search algorithms, i.e. Levy-based cuckoo search, Cauchy-based cuckoo search, Gauss-based cuckoo search and Gamma-based cuckoo search.