A study on an extended Prey-Predator algorithm

Abstract

Metaheuristic algorithms are approximate solution methods for optimisation problems which try to improve the quality of solution at hand iteratively in a random way. In recent years, various studies have been conducted in forming new metaheuristic algorithms and modifying or improving existing algorithms to enhance the performance in optimal solution search. In this study, we focus on extending an existing algorithm Prey-Predator algorithm proposed by Tilahun and Ong. Prey-Predator algorithm is a metaheuristic algorithm inspired by interaction between prey and predator among animals. The algorithm imitates the way a predator runs after and hunts its preys where each prey tries to stay with the pack trying to search for hiding place and run away from the predator. In extension of Prey-Predator algorithm, the number of both best preys and predators are increased resulting in a more reasonably exploitation and exploration so that multiple solutions can be achieved. The simulation of nmPPA is carried on ten selected benchmarks test function. nmPPA aimed to solve the problem of objective values being trapped in local optimum and to find multiple solutions at the same time.