基于两种新型遗传算子的优化组合遗传算法  

Optimization Combination Genetic Algorithm Based on Two New Operators

作　　者：帅训波[1] 马书南[1] 邵艳伟[1] 王建忠[2] 

机构地区：[1]中国石油勘探开发研究院廊坊分院地球物理与信息研究所,河北廊坊065007 [2]北京工业大学计算机学院,北京100022

出　　处：《计算机系统应用》2010年第7期98-102,共5页Computer Systems & Applications

摘　　要：针对遗传算法中全局搜索与局部搜索之间的矛盾,应用二进制编码对搜索空间描述精细、容易位值计算的特点,从矩阵遗传算子和布尔遗传算子的角度,分别对全局搜索和局部搜索的性能进行改进,并将二者组合应用,构造基于这两种新型遗传算子的优化组合遗传算法,避免了传统遗传算法中杂交率和变异率参数的选取,保证了算法的全局收敛性。实验结果表明,该算法具有更好的整体搜索性能,对应用二进制编码遗传算法求解复杂非线性优化问题具有重要借鉴意义。In order to resolve the contradiction between global searching and local searching of genetic algorithm, matrix genetic operator and Boolean genetic operator are proposed based on binary code describing search space fine and calculating fitness easier. Global searching of genetic algorithm is improved by the matrix genetic operator making good use of colony information to find new space, while local searching of genetic algorithm is improved by the Boolean genetic operator making good use of chromosome information. An optimization combination genetic algorithm is constructed through combining the two new genetic operators. It does not require parameters of crossover probability and mutation probability, which are important to general genetic algorithm. The constructed genetic algorithm is proved to be convergent, and its better searching efficiency is shown by experiments results. It is a significant reference for other optimum problem resolved by genetic algorithm based on binary coding.

关 键 词：全局搜索 局部搜索 矩阵遗传算子 布尔遗传算子 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]