# 113
Dear Wenguang Yang and Yunjie Wu
I read your paper:
New Improvement Method to Avoid Rank Reversal in VIKOR
These are my comments
1 – In the abstract you say “The phenomenon of rank reversal indicates that there is a contradiction in the consistency and credibility of the evaluation results”
How can you be so sure about this? On what grounds?
2- In page 1 “presented that rank preservation was often a normative or desired concern to make life more orderly and tractable”
In my opinion, RR is normative, it can’t be descriptive or desirable concern
3- In page 2 “Once the rank reversal occurs which means that the decision-making method itself has design flaw”
Yes, it probably is true, but not for all MCDM methods. How do you explain that there are methods that are immune to RR?
4- Page 2 “In fact, the number of indicators and the number of alternatives to be selected only depend on the evaluation problem itself, which is uncontrollable”
Agreed, and each problem is different even from another similar.
5- Page 2 “The intrinsic reason for the rank reversal is that the design of the decision-making method is always limited to the evaluation of data, which is a relative mode”
I disagree, it would mean that if in AHP, DM1 has his own data and produces RR, a DM2 with another data may be immune to RR? This is not very logical
6- In page 2 “We can get comprehensive and objective evaluation results by jumping out of the local evaluation and standing from the overall perspective”
What does this mean?
7- In page 4 “The R-VIKOR method is determined based on historical statis[1]tics for global judgment or expert knowledge”
What happens if there are not history or statistical values?
8- Page 5 “The R-VIKOR method finds the essential cause of the problem, that is, the evaluation cannot change with the change of the evaluation object”
I disagree, when you add or delete an alternative YOU CHANGE the object to be evaluated because the matrices are different. If you add you get more information about it. If you delete an alternative, you are losing information that you had before. Just as an example: Three persons are deciding about something and reach a consensus of priorities. If a fourth person joints, with new arguments, they may reinforce or not some previous data, he may or not be in agreement with what was previously agreed, and after discussion another ranking may or may be not obtained
9- In page 8 “VIKOR method is likely to cause rank reversal when the alternatives are added, deleted or replaced”
In my opinion you are mistaken because you consider the three actions on the same level. The first two effectively are in the same level but not the third. When you add or delete alternatives you gain or lose information because you change dimensional spaces, that is from 2D to 3D and vice versa. When you replace an alternative, you only change the values of a vector, but preserve the same space. The change of space has profound consequences in the ranking, and it can be easily demonstrated.
I hope my comments may be of help
Nolberto Munier
- Badges
- Science topic
- Similar topics
- Mathematical Sciences
- Graphs