A New Initialization Algorithm for Bees Algorithm

Abstract

The Bees Algorithm (BA) is a swarm- based metaheuristic algorithm inspired by the foraging behavior of honeybees. This algorithm is very efficient, simple and natural algorithm. In this paper, two natural aspects, namely the patch environment and Levy motion are employed to propose a novel initialization algorithm to initialize the population of bees in the Bees Algorithm. Thus, an improved version of Bees Algorithm is adopted based on the proposed initialization procedure. This initialization algorithm is more natural modeling the patch environment in nature and Levy motion that is believed to characterize the foraging patterns of bees in nature. Experimental results prove the effectiveness of the proposed initialization algorithm. The obtained results confirm that the improved Bees Algorithm employing the proposed initialization algorithm outperforms the standard Bees Algorithm in terms of convergence speed and success rate. © Springer-Verlag Berlin Heidelberg 2013.