Research On Multi-attribute Group Decision Making Problems Based On Genetic Algorithm

The research on multi-attribute decision-making problems has always been the focus of research in decision-making sciences.However,with the improvement of science and technology of human society in recent years,decision-making issues have become more and more complex,and many issues related to multi-attribute group decision-making have emerged.Research on such issues has gradually become the focus.In the process of making decisions,for some specific attributes,it is usually difficult for experts to use accurate numbers to evaluate the decision alternative.Generally,fuzzy language or interval numbers are used as evaluation information.When making decisions,it is necessary to consider that different attributes have different effects on the same decision alternative.At the same time,there are a number of experts involved in the group decision-making.Due to the influence of educational background,experience and other factors,different experts may make disagreeing opinions,which increases the complexity of decision-making problems.This kind of multi-attribute group decision-making problems are studied in the thesis.First of all,the basic concepts of multi-attribute decision making are introduced in the thesis,as well as its general solutions.On this basis,the theoretical basis and related knowledge of multi-attribute group decision-making are introduced,while the core problem to be solved is clarified,and the solution steps are given.Secondly,different forms of attribute evaluation information need to be transformed into the same way for later comparison.The interval number and fuzzy linguistic value are transformed into random variables,by means of introducing the concept of random dominance.The random dominance matrix is obtained by comparing two alternatives,and then normalized for the next calculation.Then,the ideal point method and the maximum deviation method are considered comprehensively,and the advantages of them are analyzed,which are combined preliminarily to solve the attribute weights.In addition,the initial alternative ranking value of each expert's decision matrix is obtained by the ideal point method.Finally,the mathematical model is established to solve the problem.The ant colony algorithm is used to get the comprehensive alternative ranking by using the initial attribute weights and the initial alternative ranking obtained.The genetic algorithm is used to optimize the attribute weights,and the comprehensive alternative ranking is re solved in the iterative process,so as to achieve the purpose of dynamic optimization.This method can comprehensively consider the decision-making alternatives of all experts in the process of iterative calculation,so that the final ranking results are more accurate,and its feasibility is verified by practical examples.