# 120

Dear Hamed Taherdoost, and Mitra Madanchian

I read your paper

Comprehensive Review of Applications, Advantages, and Limitations

My comments

1.- In page 2 you say that ANP is a new form of AHP. After 44 years it is not precisely new

2- In page 2 you say “Although AHP addresses many deficiencies of some other MCDM techniques”

And what would be those deficiencies? As matter of fact, it is the opposite deficiencies were addressed by other methods, ANP included.

The feedbacks that you mention do not exist. It was an assumption developed by Saaty, and never demonstrated

3- In page 3 “arcs are used to show the relationships in the network. The arcs’ directions impact the directional dependence here”

Not necessarily. In networks arcs normally denote precedence. If task 8 is preceded by task 4, it means that task 4 must be finished to start task 8, and it does not mean than task 4 affects task 8.

If you are in a queue wafting for service in a supermarket, the person ahead of you precedes you, but it does not mean that he can influence you. However, if a person develops some activity in 2023 that can affect his life in 2024, here there is a precedence and also a dependency., since what you do in 2024 is a consequence of your action in 2023

You can represent precedence in a network or in a matrix without problem, but how can you represent effect?

4. Page 4 “An eigenvector is used to represent the impact of each element on the other”

Based on what? There is a way, using the IF function, but I do not know how many MCDM methods can introduce this function. I least, I know one.

5. Page 4 “It can also be used to incorporate the knowledge and experience of decisionmakers in an intuitive”

No reasoning, no consultation, just intuition? Very scientific indeed

6. In Table 2. Did you realize that you enumerate advantages only from the theoretical point of view?

Is it importance for the user? No, since he is not interested on the algebra of the method.

What he wants is a method that can represent as close as possible a real-life situation, like treating several scenarios at the same time, considering alternatives characteristics as inclusivity and exclusivity, resources (where are they?), how to represent certain goals or targets?

Same for disadvantages. The user is interested in easiness in its use (ANP is too complex), speed, work load to prepare and proceeds it, etc. I have used it, and after a couple of minutes it can be frustrating. What if one a result is obtained the DM needs to change some data. Can you imagine using this method where many tests and different runs are necessary, as in research problems?

By the way, ANP is rarely seen used, in comparison for instance with AHP.

7. Page 7 “In conclusion, the ANP method is a versatile and widely applicable MCDM method that offers a range of advantages over other similar methods, such as AHP”.

This is the preferred method because the DM does not need to think, just put numbers from your mind and push the start key

You say that ANP has a range of advantages on AHP and it is true. Do you think that ANP has advantages over well-known, dependable and proved methods like PROMETHEE, ELECTRE, TOPSIS, and VIKOR? Which is the advantage of a method based on personal comparisons and deriving invented weights, which does not even consider that resources are finite, and when trying to perform a sensitivity analysis is a nightmare?

I hope these brief comments may help

Nolberto Munier