Abstract: A Self-Adaptive Artificial Bee Colony（SA-ABC） algorithm is proposed for constrained optimization problem. To make the initial colony scattered evenly on the search area, the opposite learning initialization is employed. For constraint handling, an adaptive selection strategy is designed, which can balance the feasible individuals and infeasible individuals. Furthermore, to improve the optimal ability of SA-ABC, the best-lead search equation is used in onlooker bee phase. To exam the efficiency, SA-ABC algorithm is tested on 13 well-known benchmark test functions, and the experimental results are compared with other state-of-art algorithms. The analyses of the experimental results suggest that the SA-ABC algorithm outperforms or performs similarly to other algorithms.

Key words: adaptive selection strategy, artificial bee colony algorithm, opposite learning initialization, constrained optimization

摘要： 针对约束优化问题，提出一种自适应人工蜂群算法。算法采用反学习初始化方法使初始种群均匀分布于搜索空间。为了平衡搜索过程中可行个体和不可行个体的数量，算法使用自适应选择策略。在跟随蜂阶段，采用最优引导搜索方程来增强算法的开采能力。通过对13个标准测试问题进行实验并与其他算法比较，发现自适应人工蜂群算法具有较强的寻优能力和较好的稳定性。

关键词: 自适应选择策略, 人工蜂群算法, 反学习初始化, 约束优化