小生境遗传算法修正三角模糊数互补判断矩阵一致性及排序  

Consistency improvement and priority of triangular fuzzy number complementary judgment matrix based on niche genetic algorithm

作　　者：刘胜[1] 张玉廷 于大泳[1] 

机构地区：[1]哈尔滨工程大学自动化学院,哈尔滨150001

出　　处：《系统工程理论与实践》2011年第3期522-529,共8页Systems Engineering-Theory & Practice

基　　金：国家博士后科学基金(20080430888);黑龙江省博士后基金(LBH-Z06026)

摘　　要：一致性检验、修正和权值排序作为一个整体处理,属于全局分析方法.采用遗传算法优化建立的一致性指标系数函数,引入小生境技术以保证在待优化参数较多时仍能找到最优解.分析了算法的时间复杂性和稳定性.并通过仿真对比验证了算法可在对初始判断矩阵做较小程度修正的基础上使修正后的判断矩阵具有更小的一致性指标系数.最后通过算例分析和该方法在舰船电力推进系统电磁兼容性评估中的应用,证明了该方法的可行性.该方法对克服复杂系统性能评估中人们认识的片面性具有促进意义.The global analysis method takes the consistency examination, matrix improvement and weights ranking as an integral process. The genetic algorithm was adopted to optimize the consistency index coefficient function, and the niche technique was also used to find the optimum solution when there were many pre-optimized parameters. The time complexity and stability of the algorithm were analyzed. Comparisons of simulations show that the algorithm can make judgment matrixes have smaller consistency index coefficients, which is on the basis of correcting the initial judgment matrixes in fewer degrees. The method was proved to be feasible by an example analysis and an application in electromagnetic compatibility evaluations of ship electric propulsion systems. This method is valuable in overcoming the one-sidedness of evaluating complex systems.

关 键 词：三角模糊数互补判断矩阵 小生境遗传算法 一致性指标系数 电力推进系统 

分 类 号：N945.25[自然科学总论—系统科学]