Study on the weight of evaluation index based on the rule of maximum entropy

Rationality of index weight designing is one of the most important factors which relate to whether the result for one of multi-object evaluation problem is correct or not. However, weights got from the major objective weight methods at present are all excessively highlighted separating capacity of the index, but neglect the importance and contribution for the evaluated object. Because of these above, this paper proposes the concept of "standard object", and constructs a multi- goal program model based on the maximum entropy to design weights for indexes. During the construction process of this model, we put forwards one useful supposition and two definitions, construct an evaluation matrix and carry on "equivalent" processing to the matrix. This model not only concludes the separating capacity of the index, but also includes the importance and the contribution of it. So the evaluation result of this model is much more accurate. Finally, this model is applied to a real case, and the result demonstrates that the model is a good model for solving multi-object evaluation problems.