Weight On Dea Model To Determine The Methods Of Research

Of all the system-evaluating models, DEA is one of those which have gained the most rapid development in the fields of Management Science and Operation Research in the latest decades. However, it also bears some defects. How to deal with the weights of valuation indexes is still an important unresolved problem. At present, there are many methods for the evaluation of input and output index weight in the model of DEA, which can be roughly classified into two categories. One is equal-weight-valuation which equally regards the effect of input and output index weight to the evaluating results in the traditional model of DEA, and the other is unequal-weight-valuation which is confirmed by one's preferences.Although these scholars' methods of restraining weights in the model of DEA are different, they all essentially adopt the method of subject-weight-evaluation. By this way can some problems of different essentiality be solved among the valuation indexes, yet objectivity of DEA is destroyed. The main reason for DEA's popularity is that it has two remarkable characteristics: remarkable objectivity and adjustability to the multiple-input & multiple-output system. Sacrificing the objectivity of DEA is the loss outweighing the gain. In addition, the method of subject-weight-evaluation is not suitable for empirical analysis but for normative analysis. However the method of object-weight-evaluation is just the opposite. Therefore the method of object-weight-evaluation must be discovered to remedy this defect.Grey Relational Analysis is one of the better object-weight-evaluation methods. Its essential idea is to judge the relational degree according to the resembling degree of the time sequence curve of matters or factors. If the shapes of the two curves are similar, the relational degree will be larger. That is to say, its weights are also larger. On the contrary, the less relational degree, the less weights. Thus it is unnecessary for the estimator to evaluate the relational degree of factors with his own thoughts. Therefore, it can keep DEA Model objectivity and restrain the weights to establish grey relational analysis cone in DEA model. Having established grey relational analysis cone and put it into a calculating illustration to compare with DEA model with AHP cone and to compare with DEA mode without AHP cone, it shows that the results are believable. Comparing with subject-weight-evaluation, object-weight-evaluation has two advantages. One is to have resolved the problem that subject-weight-evaluation cannot be used for empirical analysis in subject-weight-evaluation DEA mode, the other is that it has more persuasion in solving the practical problems than subject-weight-evaluation DEA model.