Enhancing Different Phases of Artificial Bee Colony for Continuous Global Optimization Problems

Abstract

The working of basic ABC algorithm depends on the functioning of three categories of bees; the employed bees, the onlooker bees and the scout bees. Although, employed and onlooker bees have different functionality, they follow the same equation for exploration and exploitation. Obviously, the performance of ABC greatly depends on single equation. In order to provide a variation in the working of ABC, we propose the use of different equations in the employed bee and onlooker bee phase. The new mechanism proposed by us for the movement of the bees depends on the convex linear combination of three candidate solutions. This scheme is initially embedded in the employed bees phase while the original equation is maintained for the onlooker bees. In the second variation the basic equation for employed bees is retained while for onlooker bees, different equation is used. The simulation results demonstrate that the modification increases efficiency and capability in terms of balancing exploration and exploitation as well as the accelerating the convergence rate of the ABC.