Experimental Study of Specific Benchmarking Functions for Modified Monkey Algorithm

This paper presents an optimization algorithm based on a meta-heuristic search inspired from the behavior of monkey. The main steps of the optimization search algorithm comprise climb process, watch-jump process and somersault process. The proposed method is a completion of the Monkey algorithm to the original steps; we have added two more: one-component perturbation and all-components perturbation. This way the monkeys have the possibility of accelerating the local convergence and also of searching better values globally. The modified algorithm for these processes have been designed and presented in the paper. The performance of the modified algorithm is tested on some benchmark functions. These functions are evaluated by considering the cases in which the problem is set as 30, 50 or even 100 dimensions. The computational results have demonstrated that the algorithm performs much better for solving a set of global optimization problems.