Abstract:

To solve the problem of evaluation index selection and weight distribution,on the basis of the idea of genetic algorithms,we propose an index system establishment method based on screening evolution model (SEM).In each screening round,the data reliability based on Euclidean distance and the index fitness based on evaluation information entropy are calculated.According to index fitness,excellent indexes are selected into the preliminary set,and inferior indexes are dealt with by crossover and mutation operations to form a new alternative set.After several screening rounds,a preliminary index set is established eventually,and the index weights are allocated and adjusted according to their fitness.The sample analysis proves the feasibility and effectiveness of the method.

Key words: index system;screening evolution model;index selection;expert assessment;weight distribution