TOPSIS Model Based On Recorded Values

Multiple criteria decision making refers to the selection of an optimal alternative among a set of feasible alternatives that cannot be replaced by each other.TOPSIS is a common method in multiple criteria decision making,which is widely used in many fields such as humanities,economy,management,etc.However,TOPSIS has its own defects,and the rank reversal has been a hot topic in academic circles,which means the increase or decrease of a certain alternative,which will lead to the reverse change of the ranking of the original alternative.The reasons for the rank reverse are mainly due to four aspects: changes in Positive Ideal Solution and Negative Ideal Solutions;changes in normalization methods;changes in index weights;changes in the relative index of each alternative.This paper mainly discusses the rank reverse caused by the change of ideal solution,the main contents are as follows:(1)In order to solve the TOPSIS rank reverse caused by the change of relative positive and negative ideal solutions,the paper illustrates that when the alternative is increased or decreased,the relative positive and negative ideal solutions are easy to change,thus changing the ranking results of alternative schemes.(2)Combined with the existing concept of absolute ideal solution,the limitation of absolute ideal solution is analyzed,that is,absolute ideal solution is not unique and definite.It is illustrated by an example,different absolute ideal solutions act on the same decision-making problem,resulting in different ranking results.(3)In view of the limitation of absolute ideal solution,this paper combines the concept of record value in the multi-attribute decision-making problem with historical data,which is to replace the absolute ideal solution with the record value to reduce the reverse in the TOPSIS and establish the TOPSIS based on the record value.The record value is selected according to the historical data,and it exists objectively.The probability of reverse order of traditional TOPSIS is calculated by stochastic simulation.The probability of reverse order of the improved TOPSIS is calculated by using the stochastic simulation method based on genetic algorithm.By contrast,the improved model can effectively reduce the reverse order phenomenon.(4)Finally,the paper compares the traditional TOPSIS and the improved TOPSIS in dealing with the rank reversal,and shows that the improved model can reduce the probability of reverse order.