运用基于GA的粗糙集方法改进经典AHP方法研究

Extended AHP Method in Virtue of Rough Sets Theory Based on GA

传统的AHP方法解决多准则决策系统有着独特的优势,但对于有些非常复杂的决策系统运用AHP进行分析后,在求解过程中,或者判断矩阵求解比较困难,或者求得的判断矩阵不能够通过一致性检验,从而使得传统AHP方法在解决某些复杂决策系统时失效.针对这种情形,提出运用基于GA的算法去约简经过AHP分析后的知识系统,然后或者直接运用RST方法中的规则进行排序而得到结果,或者对约简后的知识系统仍然用经典的AHP方法求解而得到排序结果.在此基础上,除了提出4条转换规则外,还运用该方法求解区域信息化水平比较决策系统,并与经典AHP求得结果进行比较,得出该方法比经典AHP方法更简单且可行的结论.

AHP method can be used to solve problems in multi-criterion decision systems with its special advantages. However, some complicated problems cannot be easily solved with general methods after AHP analysis, because either the complicated tasks are unable to be solved by AHP as they do not satisfy the consistency condition or the estimated matrix is too intricate to solve so that AHP is invalidated. To deal with such problems, the AHP knowledge system is reduced with the aid of GA. The issue of AHP ordering is then solved with the RST method, or problems are reduced by RST still with classical AHP method. Thus, in additiori to the four rules in changing an AHP system into an RST decision system, the comparison decision system of region informatization level is solved using this method, and the results solved are the same as that obtained by classical AHP, indicating that the method is simpler and more reliable.