Distribution in the placement of food in artificial bee colony based on changing factor

Abstract

Nature inspired technique is moderately a new research paradigm that offers novel stochastic search techniques for solving many complex optimization problems. These techniques mimic the social and natural behavior of vertebrates. The basic idea behind modeling of such techniques is to achieve near optimum solutions to the large scale and complex optimization problems which can’t be solved using traditional or gradient based mathematical techniques. In this study a recently introduced nature inspired technique called Artificial Bee Colony, which is modeled on the intelligent foraging behavior of honey bees is selected as a framework. ABC has some inherent limitations like it favors exploration in comparison to exploitation. This causes loss in domain knowledge during the successive iterations. The proposed variant is embedded with levy probability distribution and abandon factor taken from cuckoo search, to balance the tradeoff between exploration and exploitation to obtain quality food sources (solutions) as well as improves the acceleration rate. The proposed variant is named as ABC with changing factor (CF-ABC). CF-ABC is based on an assumption that the potential food sources may have different probability distributions. CF-ABC is tested and compared with state-of-art algorithms over thirteen constrained benchmark optimization problems consulted from CEC 2006 and further validated on the Software Project Scheduling problem.