摘要
针对遗传算法局部搜索能力差的缺点,模拟生物染色体中基因排列的有序性,对阀值逆序算子进行了研究,它与传统逆序算子相比,能较好地提高群体性能提高,减少了对种群多样性的破坏,改善了遗传算法的局部搜索性能,与具有全局搜索性能好的遗传算子组合,弥补了阀值逆序算子对全局搜索性能的影响,构造了一种基于阀值逆序算子的优化组合遗传算法。从理论上证明了该算法的收敛性,实验结果表明,该优化组合算法具有更好的寻优能力,对应用串型编码的遗传算法解决一般的优化问题时,具有很好的借鉴意义,阀值可根据求解问题特征和局部搜索强度而选定。
Mimicking gene order of DNA, the new conception of reverse by threshold operator is proposed to overcome the defect of genetic algorithm in local searching. Test results show the operator keeps balance by improving local searching and destroy variety. An optimization combination genetic algorithm based on the reverse operator is developed, in order to remedy destroy variety, by combining other operator with better global searching ability. The genetic algorithm is proved to be convergent. The test results show that the efficiency of the genetic algorithm in searching is better than classical genetic algorithms. Threshold is determined properly by the character of question to be resolved and the intention of local searching.
出处
《计算机仿真》
CSCD
2006年第9期175-178,共4页
Computer Simulation
基金
中国石油天然气集团科研基金资助(KF10801-1)
关键词
遗传算法
阀值逆序算子
局部搜索
全局搜索
Genetic algorithm
Reverse by threshold operator
Local searching
Global searching