# 182
Dear Caterina Caramuta, Giovanni Longo , Elio Padoano and Maria Vesela
I read your paper
Article Comparing Power Supply Technologies for Public Transport Buses through the AHP and the Fuzzy DEMATEL Method
My comments
1- In the abstract you say “The Analytic Hierarchy Process (AHP) method was used to identify the best power supply technology among a few solutions in both procedures, which differed in the number of analyzed criteria”
2- In my opinion, to compare things you need to use the same bases, otherwise you would be comparing two different problems. In all comparison problems in MCDM you need to use the same data, i.e., the same alternatives, the same criteria and the same values, as are in the initial matrix, it does not matter if a criterion applies to a certain technology or not. For instance, in a power decision problem with alternatives Wind and Solar Photo Voltaic in columns, if criterion ‘wind speed’ does not apply to PV (albeit eventually it can), just leave blank the corresponding row for solar cell in the PV column.
I believe that if the experts and stakeholders decided for a set of criteria, all of them must be used, irrelevant of their relative importance.
When the DM submits his/her results to the Board of Directors, and if one of them asked why his criterion, important for him, has not been considered, what will be the DM answer? Better no to start explaining what DEMATEL says because that director does not care. The DM must have a rational answer and the argument that there is redundancy may become a boomerang for him, because the director can ask what are his credentials and knowledge to discuss something that can be highly technical and that probably ignore?
3- In the abstract “A literature review suggested a wide set of criteria considered in the first assessment, which were then limited to the most influential criteria using the fuzzy Decision-Making Trial and Evaluation Laboratory (DEMATEL) method”
I wouldn’t do that because how do you select the most influential criteria? By intuition? Of course, there must be a ranking of criteria importance, but you cannot decide which is more important than another, for you do not know the multiple relations between them. This is the reason for using a software
4- In the abstract “Notably, the latter enabled the reduction in the number of the criteria owing to the revealing of cause–effect relationships among them”
The fact that there could be causal-effect relationships and that reduces the size of relationships does not have any mathematical support, and not either in real-life. It appears that the method proposed only attempts to reduce the task of the DM in AHP, that is an academic issue.
Of course, there could be cause and effect relationships as for instance, that many people using wheelchairs cannot access the buses
However, the most significant issue is that the paper talks about working with causal relationship between criteria, and using AHP, that does not allow that (Saaty dixit), since it only works with independent criteria. Therefore, it seems that you are using the wrong MCDM method.
5- Page 6 “The related ranking of alternatives was computed according to the preferences of the panel of experts, with reference to the priority of the criteria and performance of the alternatives”
I am puzzled by the underlined sentences. If the selection of alternatives is based on preferences of a panel of experts, why do we need a MCDM method?
6- In page 6 “Therefore, the integration of the two techniques has been meant not only to reduce the number of analyzed criteria”
It is supposed that criteria were established between the DM and the stakeholders, especially the latter, mainly following what this people wants or need, and know. If at that time they were considered important by knowledgeable officers, on what ground do you eliminate them? DEMATEL might indicate the cause-effect relationship between two criteria, something that I doubt, but a certain criterion may also have influence on others.
Page 6 “Preferences on the mutualimportance of elements are commonly expressed using the 1–9 Saaty’s rating scale”
This log scale has been copied from the Weber-Fechter psychological law that relates incentives and responses, which has been validated. Saaty said that his scale was based on this law, however, it is evident that you cannot compare a ratio analysis value with an incentive, and then assume that the response follows a log as in Weber-Fechter.
Correspondingly, in may opinion, the AHP Fundamental Table does not have a clear base. It is common in AHP to assume aspects in accordance to the method convenience, as are:
Trade- offs are equal to weights, which is not true
The resulting matrix from pair-wise comparison must be transitive. Why?
Since the matrix that indicates the DM coherence, like it or NOT, must be transitive, it applies to the real-life that also is assumed transitive. This is the most flagrant absurdity, since in addition, implies that reality is subject to what AHP says, when it should be the opposite.
In reality, I have mentioned this for years here in RG and NOBODY rebutted it. Why? Because there is no a plausible answer, perhaps you can.
7- In page 8 you say that (Di + Ri) determine the importance of a criterion while (D – Ri) indicates the mutual influence. I guess that these two definitions come from the developers of DEMATEL, but sincerely, I do not think that it is realistic, nor the conclusion of mathematical reasoning, because I do not see any reason that justify this assertion. The equation that uses DEMATEL relates Di with Ri and the builds the total relations matrix T. However, I do not understand how this matrix T reveals cause-effect relationship. For me it denotes precedence and degrees of importance, but where can be extracted that criterion c5 has an effect on criterion C2?
8- My doubts come from the causation definition, that in simple terms is an independent variable (cause) that produces a second dependent variable (effect). For instance, a lighting in a storm creates a thunder by a shockwave from air at very high temperature. I do not think that in a set of criteria, one is created by throe action of another. It does not mean that it does not exist. In a mountainous country, if there is a strong snowfall, (cause) this can paralyze car traffic (effect), in the highway. Perhaps I am mistaken, and if so, I greatly will accept a correction
9- Regarding Fig 5 you mention four key factors or criteria. In my thinking you are missing perhaps the most important: The passenger, the main actor in this problem. Of course, pax normally don;t care about what kind of power is used in the bus, however, remember that they have to pay for whatever system you adopt.
Curiously, you contemplate ergonomics of both the driver and passengers, but not mention one of the most important factors: Accessibility, that includes low-floor buses, access through ramps for wheelchairs, safety of the wheelchairs on board, someting that is common in many cities
10- The fact that the selected power is engines using fossil fuels, when all cities around the world are banning them (unless using green liquified gas) and that the social factors is at the bottom of the table, make me think that some is wrong with your methodology, and the City Hall accepted this?
These are some of my comments
Nolberto Munier