Dear Antonella Petrillo, Valerio Antonio Pamplona Salomon, Claudemir Leif Tramarico
I read your paper
State-of-the-Art Review on the Analytic Hierarchy Process with Benefits, Opportunities, Costs, and Risks
These are my comments
1- In the abstract you say “Aggregation approaches and outranking approaches are better classifications”
I agree with this classification better that “American vs. European schools”. For your information there are methods that apply both.
2- In page 2 “The choice of an MCDM method should be based on characteristics of the decision problem”
I also agree with this, but unfortunately, practically in all MCDM methods, some characteristics are ignored in the modelling due to the inability of methods to cope with them. For instance, resources and limitations, inclusive and exclusive alternatives, precedence, time, binary variables, etc.
In my opinion, the choice of a method is simply: Choose the MCDM method that best adjust to the characteristics of your problem.
3- “One main reason for the AHP’s leadership in MCDM is its solid mathematical foundation”
This is inexact. AHP does not have any mathematical foundation, except in the use of Eigen values.
Let’s see, why I say this. Do you think that there is mathematical foundation by:
a) Using pair-wise comparisons? No mathematical supportand it is a highly criticized procedure.
b) Assigning values to criteria based on intuition? Is this scientific, and what happens if other DM thinks different?
c) Accepting that the final decision of the DM is controlled by a formula, and forcing the DM to correct her/his own estimates? So, a formula, to get transitivity, supersedes the honest findings of the DM,
d) Assuming that criterion trade-offs are equivalent to criteria weights? These are two different concepts
e) Assumming that what is in the mind of the DM is applicable to the real-life, and thus accepting that it is also transitive? What kind of mathematics supports this?
f) Using a logarithmic table, the ‘Fundamental scale’, based on the Weber and Fechter laws, on stimulus and results, and then AHP comparing invented weights to stimulus?
The Dictionary defines stimulus as “Physiology, Medicine/Medical. something that excites an organism or part to functional activity”
Not even a remote relationship with the ‘weight’ concept.
g) AHP is unable to deal with complex scenarios, because its rigid lineal hierarchical structure that cannot represent transversal relationships.
Some AHP drawbacks were refuted by Dyer in the 90s. and that Saaty responded, but nothing can be extracted from those rebuttals. To be fair, rank reversal was discovered in AHP, but it is present in all MCDM methods, not only in AHP
4- You talk about BOCR as it were something new, when it started in the 50s, when the old C/B analysis was considered no longer appropriate.
Why the four criteria BOCR are mutually exclusive? Normally they are considered in the set of criteria. MCDM is not looking for optimality,since normally, it does not exist. All MCDM methods look for a balance between opposite criteria like B and C.
Exclusivity means that BOCR cannot be together and this is not realistic, nor practical, because it is a common feature in most scenarios. If you want more information, I will be glad to supply examples, albeit not using AHP
You are mistaken. A criterion can be used twice, for instance, a criterion asking for minimization, and the same criterion, with the same values, asking for minimization. I use it frequently. The software must find the equilibrium between those extreme values.
You talk about ‘important criteria’? And how do you select those criteria? Just by the weights values? There is not a mathematical support for that. It is intuitive, no more than that.
And what if there is redundancy? Which is the effect? From the mathematical point of view, none.
You are referring to AHP but at the same time make references to ANP.
There is a large difference, since the ANP structure is able to handle complex scenarios because it works with a network. Probably Saaty developed it reckoning the limitations of AHP.
5- Page 3 Figure 1. Sorry, but you cannot apply AHP to this problem; AHP theory explicitly says that all criteria MUST be independent, which is not the case in your example, quite the opposite, there are many transversal interrelationships.
6- In page 4 “First, this alternative may be too risky compared to alternatives one and two”
Obviously, you do not consider that an alternative may be too risky, but also it may have some properties that compensate for this risk.
I am not referring to the compensation issue used in weights. The problem with AHP and other methods, is that elements of the decision matrix are considered in isolation, when in reality, according to System’s Theory and reasoning, they have to be considered as a whole, holistically. For instance, you can reduce risk by increasing costs or/and decreasing benefits. Therefore, you have to consider both at the same time.
I hope my comments may help
Nolberto Munier
Nolberto Munier
Thank you for your detailed feedback and insightful observations on the "State-of-the-Art Review on the Analytic Hierarchy Process with Benefits, Opportunities, Costs, and Risks" by Antonella Petrillo, Valerio Antonio Pamplona Salomon, and Claudemir Leif Tramarico. Your perspective sheds light on critical aspects of the Analytic Hierarchy Process (AHP) and the broader context of Multiple Criteria Decision Making (MCDM) methodologies. Allow me to provide some clarifications and responses to the points you raised:
Your insights provide valuable perspectives for further examination and exploration of the Analytic Hierarchy Process and its implications within the realm of Multiple Criteria Decision Making. I appreciate your engagement with the subject matter and your dedication to advancing the understanding of decision-making methodologies.
Sincerely,
Dear Qamar
NM- On May 11, 2023 I started reading and commenting RG published papers on MCDM, and today or tomorrow I expect to publish in RG the hundredth analysis and comments. All of them are in RG.
It has been a very nice surprise to receive you comments, and I thank you for that.
Do you know how many responses I have received out of 99? Only 2
The first, months ago when the author withdrew the paper based on my observations.
The second? YOU.
Refusals and feedback from authors? None
Of course, there is no obligation to write or feed me back, but what really surprises me is that NOBODY defended or backed what he/she wrote, even when normally there are several authors. I believe that if a person writes something and somebody else refutes the writer, the normal procedure would be to answer and refute it or not. This happened in the 90s when Dyer commented on AHP, and Saaty and Harker published their refusal. The famous letters from both sides are in the Web
Nobody asked me to write and publish my comments on work of others, and I don’t ask or expect anything, just my desire to collaborate in the improvement of this discipline. I do it because I have dedicated the last 20+ years in studying and analysing MCDM, a fascinating subject for me, and in my opinion, this discipline is not pursued correctly. There are more that 200 MCDM methods, and all of them, are unable to address complex scenarios, with the possible exception of the grand dad of all methods, Linear Programming, created about 1940.
Qamar UI-Islam- Thank you for your detailed feedback and insightful observations on the "State-of-the-Art Review on the Analytic Hierarchy Process with Benefits, Opportunities, Costs, and Risks" by Antonella Petrillo, Valerio Antonio Pamplona Salomon, and Claudemir Leif Tramarico. Your perspective sheds light on critical aspects of the Analytic Hierarchy Process (AHP) and the broader context of Multiple Criteria Decision Making (MCDM) methodologies. Allow me to provide some clarifications and responses to the points you raised:
1- Classification Approach: I appreciate your agreement with the aggregation and outranking approach classification, emphasizing their relevance in the context of AHP and MCDM methodologies.
NM - AHP was a method created in the 50s., using a concept never employed before in decision-making, and that at said time worked well, because it uses the hierarchical structure that was been employed by enterprises since the Egyptians, or before. That is: “Do as I say, don’t argue”
Saaty developed his method when he was working for the US Army, and possibly adopted its hierarchical lineal and rigid structure for AHP.
Unfortunately, the method incorporates aspects that border the absurd when applied to MCDM, like the pair-wise comparison, developed by psychometrician L. L. Thurstone, in 1927.
You can apply it to decide which restaurant to go for dinner, or the place for your vacations, or, in the case of the military, considering different strategies and war equipment, but not in a problem like selecting the best location for an industry. Here, there are multiple aspects, all of them related, and on what the DM does not have any control, let alone authority.
Qamar UI-Islam- Choice of Method: Your insight into the selection of an appropriate method based on the problem's characteristics aligns with the practical considerations of matching the method to the specific nature of the decision problem.
NM - It is not mathematics, only common sense, like you don’t buy a suit if it does not fit you
Qamar UI-Islam- Mathematical Foundation of AHP: Your critical assessment of the mathematical underpinnings of AHP raises pertinent questions regarding the method's reliance on subjective pair-wise comparisons and the practical implementation of criteria trade-offs.
NM - I already commented about pair-wise comparisons. Regarding trade-offs, considered as criteria weights, it was only a Saaty assumption,as he described it in one of his writings. But trade-offs and weights have completely different meanings and use. The first one is a balance, while the second is a quantification of criteria. The fact that these weights, DETERMINED WITHOUT CONSIDERING THE ALTERNATIVES, is a fallacy. It can be easily demonstrated using entropy weights.
1. Qamar UI-Islam- BOCR Criteria: Your emphasis on the dynamic nature of criteria relationships and their potential for non-mutually exclusive interactions contributes to the discussion around the practical applicability of BOCR criteria and their role in decision-making processes.
NM- Thank you for your appraisal
2. Qamar UI-Islam- AHP Application Challenges: Your detailed evaluation of the constraints and limitations of applying AHP to complex decision problems underscores the need for a more comprehensive and adaptable approach, especially in scenarios with interconnected criteria.
NM- In reality, there are more than 100 researchers that addressed these issues in AHP, who I mention and identify by name in one of my books, and only a handful that are in favor of the method, also identified; therefore, they are not only my ideas; what I did is to enlarge and illustrate these fallacies
Regarding criteria interconnections, there are very few scenarios where criteria are independent. What really surprises me, after reading hundreds of published papers that ‘solve’ this type of problems using AHP, even in reputable journals, is that reviewers accept them. WHY?
By ignorance, negligence, or vested interests, because some reviewers are only interested in being mentioned as reviewers of a journal and mentioned in ‘Publons’, which is good for their CV?
It is impossible to ignore essential points like a primordial condition in AHP, clearly established by Saaty, about the necessity to have only independent criteria, or trade-offs equivalent to weights, and other evident misconceptions.
3- Qamar UI-Islam- Risk-Benefit Trade-Offs: Your emphasis on the interconnectedness of risks, benefits, and compensatory mechanisms within decision matrices highlights the significance of holistic and systems-based considerations in decision-making frameworks.
NM- Again, it is only common sense. We are not talking on Quantum Mechanics
Qamar UI-Islam -Your insights provide valuable perspectives for further examination and exploration of the Analytic Hierarchy Process and its implications within the realm of Multiple Criteria Decision Making. I appreciate your engagement with the subject matter and your dedication to advancing the understanding of decision-making methodologies.
NM- Thank you for your words, which I deeply appreciate. Maybe this communication may encourage other researchers to publish and discuss their concerns and why not, support for AHP and other MCDM methods.
Regards
Nolberto
- Badges
- Science topic
- Similar topics
- Mathematics
- Algebra
- Matrix