Abstract: To overcome shortages of the method of multiple factor decision making with variable weights(MFDMVM), a new weights implicitly changing method for multiple attribute decision making (MADM) is presented based on the principle of analytic hierarchy process(AHP) modified by Belton and Gear (shorted as B/G-AHP). With the modeling method, one new approach to MADM with weights implicitly changing is presented. It has three advantages over the MFDMVM. First, the needed preference information to weights changing is directly provided by the decision maker (DM), and thus the approach can well overcome the decision analyst’s subjective arbitrary influence and better reflect the DM’s real preference. Second, it need not subjectively transform attribute states into preference on attribute states, thus it avoids error interrupts resulted from the state-to-preference transforming. Third,unlike the MFDMVM, it has adopted the optimization technique to reduce the influence of errors within the DM’s subjective judgments. Numerical analysis shows that the new approach does have the efficient capability of giving variable weights to different decision alternatives, can give more acceptable evaluation results than the MFDMVM, and is more applicable to real-world decisions.

Key words: multiple attribute decision making, implicitly-changing weights, decision analysis, analytic hierarchy process