Combination Weighting Method Based on Maximizing Deviations and Normalized Constraint Condition

When the weight of each attribute is determined in the multiple attribute decision making problems, calculated by the method of subjective values or objective values solely will cause the problem that weight coefficient is not reasonable. So the paper puts forward the weightingmethod which is based on maximizing deviations and normalized constraint condition. The method integrates the subjective and objective weighting information. On the one hand, the deviation between each weight vector which is determined by the various weighting method makes the maximum of its total deviation. On the other hand, the various evaluated object integrated value makes the maximum of its total evaluated value. Thus we establish a double objective optimization model. What's more, we deduce the weight calculation formula by solving the model. Finally we have an experimental analysis. It proves that the combination weighting methodcan reflect the relative importance of each indicators and the information that index itself contains. In other words, it can reflect the subjective and objective decisions which make the weighting results more reasonable.