A Research And Application Of Multi-attribute Combined Decision-making

As one important branch of decision theory science,the Multi-attribute Decision-Making problem is widely used in various domains such as society,economy,management and military.The differences in different decision-making problems issues,the complexity of the decision information obtained and the application of decision method are coming along with the popularity of applications.The problem caused by these factors is that the stability and reliability of the evaluation results are poor,conclusions drawn from a single evaluation method shows inconsistent.Therefore,in order to improve the closeness between the evaluation results and the real results,people’s research focuses on how to adopt appropriate decision-making methods to improve the stability and reliability of the evaluation results,“Combination decision” came into being.However,the research on multi-attribute combination decision theory and method is still not perfect,and further research is needed.Therefore,this paper uses a comprehensive analysis and application way of thinking to study the problem of multi-attribute combination decision making,combined with operational research,system science,deep learning and computer science.The main research questions and results of this paper were divided into three parts:(1)This paper summarized,sorted out and analyzed the existing multi-attribute combination decision making methods.Firstly,the different research emphases of combinatorial evaluation at home and abroad were summarized through the data of journal website.Then in order to select the appropriate combination method set,this paper composed the required method set according to the characteristics of the method and special rules from different multi-attribute decision methods.Finally,this paper combined the results of methods in the method set to obtain the combination of method results.(2)This paper studied the combined results produced by different methods.To solve the order preserving problem of multi-method evaluation results,this paper presents a method of result combination evaluation with optimal sequence preserving feature combination.A method was firstly provided to establish a Hadamard matrix of combined evaluation consequence and each single evaluation method.Then the optimal combined weight coefficient was obtained by window width function.Amethod of designed and used the average deviation was proposed to measure the result of this approach.An instance was used to compare this approach with Fuzzy-Borda and Maximizing-deviation and the result showed a better consistency between this approach and other single evaluation method.(3)Aiming at the fuzzy multi-attribute combination decision making problem,this paper studied the fuzzy set and membership function,and gived full consideration to the non-membership degree and hesitation degree of fuzzy set.This paper proposed an improved scoring function,which combines the intuitive fuzzy number and ARAS method to evaluate the evaluation object.According to the characteristics of triangular fuzzy number,a method of attribute reduction combined with dominant rough set DRSA was proposed to determine the classification basis,and the method was applied to the evaluation system of enterprise financial status.Through analysis and comparison,the validity of the method is verified.