Authors: Lin Cheng; Hailian Dong; Qingzhen Zhang; Zhenghong Liu
Addresses: School of Automation Science and Electrical Engineering, Beihang University, Beijing 100191, China ' Beijing Shengfeifan Electronic System, Electronic System Technology Development Co., Ltd., Beijing 100141, China ' School of Automation Science and Electrical Engineering, Beihang University, Beijing 100191, China ' Beijing Polytechnic, Beijing 100176, China
Abstract: An adaptation of classical artificial bee colony (ABC) algorithm based on imitating the foraging behaviour of honey bees is presented for constrained numerical optimisation problems. The modifications focus on improving the operator of candidate food sources updating by using a quantum delta potential well model. The well model described the behaviour of bees in a quantum multi-dimensional space and realises quick convergence of algorithm because of available food sources information utilisation. Furthermore, two dynamic tolerances changing in exponential form are introduced to help the honeybee colony converge around the feasible region. Finally, a general mechanism of selection probability which associates with the fitness of food source is proposed. The new algorithm called QABC is tested on a set of 13 benchmark constrained non-linear optimisation problems (CNOPs) and the comparison against the original algorithm and some state-of-the-art algorithms gives the reasons for the modification.
Keywords: ABC algorithm; artificial bee colony; quantum delta potential well model; dynamic tolerance; constrained optimisation; nonlinear optimisation; ABC adaptation; quantum behaviour.
International Journal of Service and Computing Oriented Manufacturing, 2016 Vol.2 No.1, pp.50 - 66
Available online: 22 Mar 2016 *Full-text access for editors Access for subscribers Free access Comment on this article