基于克隆选择的小世界优化算法

摘要

针对小世界算法在多极值等复杂函数优化中存在算法后期种群多样性退化、全局搜索效率下降等问题,提出一种基于种群克隆选择的小世界优化算法.该算法以小世界现象信息传递的高效性改进克隆过程中体细胞高频变异的随机性,实现克隆增殖、克隆选择以及小世界网络短连接等算子在局部空间的搜索,克隆删除与小世界随机长连接在全局空间的搜索.实验结果表明:各种克隆算子与小世界变异算子相结合,增加了种群的多样性,扩大了搜索范围.与其他算法相比,该算法在收敛速度和多极值点函数搜索能力等方面具有明显改善.%To avoid trapping into degradation in diversity of population latterly and improve global searching efficiency in multi-extreme function optimization, a small-world optimal algorithm based on clone selection was proposed. Small world phenomena, known for high-efficiency of information flowing, were introduced to change in the randomness of somatic cell mutation. The clone proliferation, clone selection and small-world clustering were used to achieve the local searching; the clone deletion and small-world long-range shortcuts operators were combined to enhance the capability in global searching. The simulation results show that the combination of clone operators and small-world mutation operators can improve the diversity in population and expand the scope of search. Compared with other small-worEd algorithms, the proposed algorithm has remarkably improved the converging rate and the search ability in multi-extreme functions.