# 103

Dear Igor pinheiro de araujo costa , Isaque david pereira de almeida , Miguel angelo lellis moreira , Adilson vilarinho terra , Sergio mitihiro do nascimento maeda , Carlos francisco simoes gomes,and Marcos dos santos

I read your paper:

Study for decentralization of naval power in the Brazilian navy: structuring and mathematical modeling in the light of the electre-mor method

My comments

1- In page 5 you say: “Also, the ELECTRE-MOr method presents two ways of obtaining the lower limits of pre-defined classes, providing two distributions of the alternatives”

Which are the pre defined classes? Distributions of alternatives? Do you have probabilistic values for them?

2- Page 5 “According to Souza et al. (2018), showing two ordinations after applying the method allows a sensitivity analysis of the result, giving greater transparency and robustness to the decision-making process”

The objective of the sensitivity analysis is to check the robustness of the best solution found, not of the process

3-Page 7 “SAPEVO-M method (Gomes et al. 2020), transforming ordinal preferences of criteria into a vector of criteria weights and integrating the preferences of different DMs, by applying the scale of pairwise comparisons presented in Table 3”

It appears that the SAPEVO method simply turn a linguistic language into crisp values, and since the preferences are personal, the resulting values are invented, without any mathematical support. Therefore, another DM or DMs group may come with different values. Is this science?

In my opinion this is an unreal approach of a complex problem. Table 5 shows only 7 criteria, which is from my point of view, clearly not enough. Nobody can solve this type of problem with only 7 criteria..

These are my few comments because I believe that there is with no need to continue. I hope they can help

Nolberto Munier