# 102

Dear Tran Van Dua

I read your article:

Development of a new multi-criteria decisionmaking method

My comments:

1- In the abstract you say: “That difference is reflected in the fact that when applying this method, the decision maker does not need to normalize the data nor determine the weights for the criteria”

Effectively, since everything is reduced to three integers 1, 0, -1, it is obvious that you don’t need normalization, since they are no units of measure.

However, you give the same number to a difference of say 3 and 8 than another between 0.5 and 4500. I don’t think that this is correct.

2- You say “In each example, the result of ranking the alternatives using the CURLI-2 method has been compared with those using other different MCDM methods.”

This comparison is irrelevant, it does not have any meaning and there it could also be due to coincidence

3-“The best alternative determined when using the CURLI-2 method always coincides with the use of existing MCDM methods”

This sounds strange, because this normally does not happen with other methods, when different MCDM methods, addressing the same problem, yield different results, including the best alternative, and you state it explicitly. I don’t understand why CURLI-2 is the exception, although could be, if you explain its algorithm.

4- In page 1 “Thus, it can be seen that when using the method of group four, the decision maker will eliminate the difficulties in data normalization as well as determining the weights for the criteria”

Normally, you may eliminate normalization, but not criteria weighting because that means that all criteria have the same importance, and in general each one has a different importance.

5- In page 3 “The result of ranking the alternative done by this method will be used to compare to the ranking result done by CURLI-2 method”

And what these comparisons prove? Nothing

6- In page 4 “Add the scoring matrix for each criterion”

In my opinion, computing each criterion independently and further adding up results is incorrect. Why?

Because the decision matrix represents a system where everything is related to something else. Therefore, you can have criterion C6 that receives input from criterion C2. How do you compute the contribution of C2 on C6?

I hope that these comments can be of help

Nolberto Munier