Abstract:By simulating the process of human thinking

a solving method for multi-attribute group decision making problem based on granular computing was proposed. Firstly

mathematical description of granular computing structure model was discussed. Secondly

in order to give a fine description to single decision maker’s thought at different granular layers

granular information entropy measurement under relative meaning was given. Through definitions of matching rate and coverage rate under granular divisions

the optimization of granular layer based on similar threshold was carried out

and the final optimization result

which indicates single decision maker’s decision thought

was exhibited by weight values of feature attributes. Thirdly

based on values of weight vectors which indicate decision makers’ decision thought upon different granular layers

the nonlinear optimization model was built up to rebalance all decision makers’ opinions at different granular layers and get the final optimal weight values of feature attributes

which can be recognized by all decision makers. Thus

an integrated and qualitative problem solving environment for multi-attribute group decision making was reached. Lastly

a graduate student admission interview assessment example was given to prove its feasibility and superiority.