Self-adaptive Artificial Bee Colony with role transformation for engineering constraint optimization problem

Abstract

In order to solve engineering constraint optimization problems, A self-adaptive artificial bee colony (ABC) with role transformation is proposed. The translation function of unemployed bees' location is improved, the unemployed bees focus on the location of the employed bee and search in high speed, and the algorithm is avoided trapping into local optimum due to the method. A self-adaptive method is improved to assign to initial evolutional generations of unemployed bee's population and to reduce parameter settings, the convergence speed and precision of ABC is improved effectively. The algorithm is used for two standard engineering design problems, Tension/compression string design problem (T/CSP) and The welded beam design problem (WBP). The results indicate that self-adaptive artificial bee colony with transformation has faster convergence speed, higher precision, stronger solidity and better ability of global optimization.