Dear Wojcieh Salabun and coauthors: Bartłomiej Kizielewicz , Andrii Shekhov
Reference is made to your article “ pymcdm—The universal library for solving multi-criteria decision-making problems"
I read it and my comments are:
1- In the an abstract you say “comprehensive software that should have relevant components”
What is the meaning if this sentence? Which are relevant components?
If for comprehensive software you mean a MCDM that allows inputting some characteristics or ‘components’ that belongs to a project, I plenty agree with you. For instance, at the beginning, PROMETHEE did not consider resources, and after a time they were incorporated, the same with GAIA, and thus giving a realistic graphic vision of the result. The same for AHP, that was greatly improved by ANP by considering any type of relationships. And of course, Python is probably the best language for MCDM, but it can’t invent procedures that don’t exist. For instance, it can’t solve multiple scenarios if the method does not contemplate that characteristic that a problem may have.
As a matter of fact, I have been fighting for this subject during years because present-day methods can’t incorporate aspects like precedence between alternatives, or incorporate binary notation that allows representing inclusivity and exclusivity, or than consider the time element in a portfolio of projects, etc.
2- You say “Therefore, this paper proposes a flexible library written in Python 3 for multi-criteria analysis/decision-making”
I agree that are many characteristics as those mentioned that can be solved by some of the methods in the library. However, for most semi-complex and complex problems I don’t think that there is in the library a method able to solve them. By the way, which is the difference between this library and ‘Decerns: A Framework for Multi-Criteria Decision Analysis ‘(Boris Yatsalo), that has been around for some years now?
3- You say in page 2 “Therefore, simple techniques may not be enough when evaluating decision-making options”.
I agree in a 100% with his statement. Now, my question: If we know that lack of capacity, why are we still using anachronic methods and don’t develop new ones.? This has been partially done by addressing aspects such as rank reversal with new methods like Spotis.
4- Very good your comments in page 3, that briefly and precisely define the characteristics of each method. It can be very useful to practitioners that are confused and lost because the large number of different MCDM methods.
5- The paper says ‘flexibility and accuracy’. How does it mean ‘flexibility’? For what?
How can you measure accuracy?
6- You say “The continuous development of multi-criteria decision-making/analysis methods also creates a gap related to the lack of tools to use them”
Absolutely true, and in my opinion, this is the most important challenge for these methods
7-You say “Moreover, projected tests based on sources from the literature ensure the reliability of the library and the outputs it provides”
I disagree, for there is nothing that can ensure reliability of the output, whatever the method used.
8- In page 7 you say “Studies show a fundamental difference in the results they receive”
This old problem is, in my opinion, produced by the subjectivity of diverse methods, due not only to subjective weights but also to different assumptions, most without any mathematical support. I believe that one way of avoiding this disparity is working with real values without any distortion due to weights (except using objective weights), as well as without assumptions and thresholds. Since for a certain scenario all methods start from the same decision making, aim at the same goal and use mathematical algorithms, logic indicate that all results must coincide.
The DM input, very important and fundamental, can take place one a result, based on reliable numbers, is obtained.
9- You say “The tools that exist to date mostly center around a few selected approaches, sometimes insufficient for a comprehensive analysis of a given problem”
Excellent definition of the main problem in MCDM!
10- On page 8 you say “Sometimes it is difficult for the decision[1]maker to determine to what extent the criteria are relevant to the problem at hand, either under their complexity or the expertise required”
Determining criteria for a certain scenario is indeed difficult if the DM relies in his/her own expertise, or lack of it, because in this circumstance he must have the knowledge of an engineer, an accountant, a social worker and an environmentalist, a geologist, etc., at the same time.
Therefore, what he must do is to consult all stakeholders involved in the project. It is from them, that the DM can obtain the information he needs, and it does not matter if it is contradictory or redundant, just let the software to solve this problem. If as an example, the production manager puts emphasis in needing some especial equipment to be considered, and the financial officer says that there is no money for it, both positions must be taken in to account, because IT IS REALITY.
It is not for the DM to decide who is right and who is wrong, or what is valuable and which not.
Reading many papers published, it is possible to check this fact, however, they are approved by reviewers.
I hope that my comments help
Best regards
Nolberto Munier
Dear Nolberto Munier You comment on the open platform of RG and not to the authors directly. You refer to: Article pymcdm—The universal library for solving multi-criteria deci...
Just contact them directly (and possibly refer to this discussion) to ensure that they are aware of your attempt to discuss this paper.
Best regards.
Dear Robert Keller
Thank you for your message, but could you please explain the reasons for your position?
If we were talking about private conversations, I will agree with you, but this not private, this is public, since it is a forum.
I am just using the opportunity that RG offers for open discussion, when says. “Please provide details or a starting point so others can contribute to the discussion”
Am I doing anything unethical? There is even a movement that claims that what journal’s reviewers say, should be published, together with the approved paper.
If somebody puts an article on RG, I understand is for everybody to read, learn and make comments, which is the objective of a discussion, of course, in an educated and professional manner.
I don't criticize, judge ,or act as a reviewer of an article. I only read it and give my opinion, something that anybody can make. What I do, is to put my experience at the authors ‘service as well as for readers.
Can I have errors in my comments? Of course, and I will be glad that if somebody publicly rerbutts me
In my view, and since always justify my assertions, many times using examples, I understand that my comments are helping people to be aware of different aspects of MCDM, considering the large variety of subjects addressed.
If addressing my comments just to the authors, and I am talking in general, not referring to the authors of the article you mention, most of the times they don't even acknowledge reception. Along my many years in RG, I have asked and encouraged in uncountable opportunities, to be openly confronted, rebutted and contradicted, if necessary, about what I publish. In about ten years, and after hundreds of messages, I NEVER received an answer.
Not so long ago, I commented an article, finding things that were flagrantly inexact. The author publicly acknowledged the error and withdrew the article.
I believe that we must not use this magnificent platform as a private exchange of opinions, and don’t think either that was the intention of its creators. I am making comments on articles content, not on the authors.
If something is wrong, it must be said publicly, in the open. If there is a conceptual or a mathematical error and it is communicated privately to the author, he/she may or not correct it, but the people that already read it, continue accepting something that was wrong, which could induce to their own errors.
Regards and thanks your message
Nolberto Munier
Dear Nolberto Munier There seems to be some misunderstanding here. All I wanted to make clear is that you addressed your question to the authors, under the assumption that they will read this. Since there are countless questions and discussions here on RG they will never see this unless they follow you or use the right keywords in the search box. I just wanted to make clear to you that it is better to address them directly to ensure they will.
Best regards.
Dear Robert Keller
I really appreciate your answer, and indeed, it appeared that there was a misunderstanding. Thanks for your good intentions.
Your point is certainly valid; there is a chance that the authors don’t read my comments in RG, but I guess that this is uncommon, because my commentaries are in the same site where they published their article, and usually there are several authors. It must be unusual that all of them choose not to see what is happening in ResearchGate, but, of course, there is always a probability of this not happening.
My contributions have a double purpose: 1) To alert an author that in his/her article there is something inexact or erroneous, and 2) to warn people who read the paper, that there is something incorrect, if they decide to replicate a certain procedure from the original.
If the authors, for whatever reasons, don’t see the comments, for sure there will be some people that do. As a matter of fact, the specific paper you referred to, has been read by 20 people (RG information), within the first two days of publication. Other articles are read by only one person, which is enough to justify my effort, and another by hundreds.
Therefore, knowing that I could help 20 people not to make the same error as the original authors innadvertedly did, is for me a proof that the discussion system works, and a reward for my efforts.
Once again, thank you for your interest
Best regards
Nolberto Munier
- Badges
- Science topic
- Similar topics
- Geoscience
- Cartography
- Mapping
- Maps