Abstract: A composite artificial bee colony algorithm is proposed for constrained optimization problem. The multi-dimensional random mutation operation and optimal guided mutation operation are introduced to balance the exploration ability and exploitation ability of the algorithm. The[ε]constraint and feasibility rules are combined to balance the objective function and constraints, which can accelerate the convergence of the algorithm. Experiments on 20 benchmark test functions in CEC 2006, 18 benchmark test functions in CEC 2010, and 3 practical engineering optimization problems show that the algorithm is feasible and effective for constrained optimization problems.

Key words: composite artificial bee colony algorithm, constrained optimization problem, [ε]constraint, feasibility rule, engineering optimization

摘要： 针对约束优化问题，提出一种复合人工蜂群算法。该算法引入多维随机变异操作和最优引导变异操作平衡算法的探索能力和开发能力。将[ε]约束和可行性规则相结合平衡目标函数与约束，加快算法的收敛。通过对CEC 2006中20个测试函数和CEC 2010中18个测试函数及3个实际工程优化问题的实验结果分析表明，该算法对约束优化问题可行有效。

关键词: 复合人工蜂群算法, 约束优化问题, [ε]约束, 可行性规则, 工程优化