Competition over Resources: A New Optimization Algorithm Based on Animals Behavioral Ecology

In the recent years, many heuristic optimization algorithms have been developed. A majority of these heuristic algorithms have been derived from the behavior of biological or physical systems in nature. In this paper, we propose a new optimization algorithm based on competitive behavior of animal groups. In the proposed algorithm, the whole population is divided into a number of groups. In each group, the best searching agent spreads its children in its owned territory. Any group which is not able to find rich resources will be eliminated form competition. The competition gradually results in an increase in population of wealthy group which gives a fast convergence to proposed optimization algorithm. In the following, after a detailed explanation of the algorithm and pseudo code, we compare it to other existing algorithms, including genetics and particle swarm optimizations. Applying the proposed algorithm on various benchmark cost functions, shows faster and superior results compared to other optimization algorithms.

I. Introduction

In the recent years, meta-heuristic algorithms inspired from behavior of nature phenomena are becoming powerful methods for solving optimization problems. Classical optimization algorithms do not provide a suitable solution for problems in high-dimensional search spaces. Over the recent decades many heuristic algorithms have been invented and have become increasingly popular [1]. Examples of notable heuristic algorithms are ant colony optimization [2], particle swarm optimization [3], artificial bee colony [4], and some more recent algorithms, like firefly algorithm [5], cuckoo search [6] and the bat algorithm [7]. The available algorithms are extensively used in different optimization problems, such as industrial planning, resource allocation, decision making, machine learning, etc. [8]–[11]. These algorithms solve various optimization problems, however there is no specific algorithm to achieve best solution for all optimization problems. Thus searching for new evolutionary algorithms is an open topic.