摘要

A multi-members artificial bee colony algorithm (Multi-ABC) is proposed for constrained optimization problems. For enhancing the search capability in the feasible region, Multi-ABC uses a multi-members mechanism. Then, to enhance the diversity of the population, Multi-ABC introduces a new selection rule. In addiction, when dealing with the problems with the equality constraints, Multi-ABC uses a dynamical violation to increase the probability of finding feasible region. Multi-ABC is tested on thirteen well-known test problems. The results demonstrate good performance of Multi-ABC in solving the problems with linear equality constraints or with optimum in feasible domain, when compared with modified ABC algorithm.