Dear Carlos Francisco Simões Gomes, Francisco Sousa, Teresa Pereira, Marisa Oliveira

I read your paper

Multicriteria Methodology for Selection of a Personal Electric Vehicle

My comments

1- Abstract, Yes, I agree that PROMETHEE is powerful. AHP? It depends of what you call powerful

2- In page 3 you say “The AHP method was used to define criteria”

No MCDM method designs criteria, for they are selected considering the characteristic of the problem

3- Page 4 “An MCDA can be defined as a set of techniques that are designed to search for several alternatives within multicriteria and conflicting objectives”

I would say that MCDA determines the best alternative of a set of given alternatives subject to a given set of criteria

4- In page 5 “The principle of this method is to break decision problems into smaller, understandable parts, analyse each part separately, and then integrate the parts in a logical way. MCDA provides support for identifying components of a decision-making problem”

This is true for some methods but not for the most. In reality this procedure violates a basic principle of system theory that considers a whole problem as a system that can’t be solved by parts. Normally, the result is greater that the sum of the parts.

5 – Page 5 “The GAIA plane is unique in that it projects both the alternatives and the criteria in the same space”

This is incorrect, Linear Programming does it since the 50s. You can represent in the same space, activities, criteria and even get the result (in elemental problems), since complex problems, with many dimensions cannot be represented in our world. GAIA is an excellent procedure and what it does I is visualize the results of a problem, and determining relative importance of criteria.

6- Page 6 “he AHP is one of the most well-known MCDA methods for ranking the various alternatives for achieving a goal. It is based on the foundation of adding weights to the attributes”

It is true that AHP is one of the most well-known MCDA methods. the. According to you It is based on the foundation of adding weights to the attributes.

It does not add weights to the attributes;it assigns weights to criteria which have different attributes.

Apparently, you equalize attributes and criteria, and they are not the same.

7- Page 7 “The problem gets more complicated in the PROMETHEE method when the number of criteria exceeds seven”

I don’t know where you got that concept. PROMETHEE can deal with many more criteria than seven, without any problem.

8 – Page 7 “There is no tangible method for calculating weights in PROMETHEE This process can be done analytically in AHP.”

No MCDM method produce its own criteria weights, except AHP and ANP. You can use objective weights that are exact, from statistics or entropy, why to work with invented weights from AHP?

9- According to the Dictionary analytical means “relating to or using analysis or logical reasoning”

Where in AHP, based on intuitions, do you use analysis or logical thinking?

10- “The results of the PROMETHEE can be explained with high visuality. Therefore, the effect of each criterion on the result can be better understood. The GAIA technique also helps visualising the results.”

Agreed

11- Pag 8 “The criteria "Price" and "Autonomy" have the highest impact on the goal”

Both criteria are related because price is related to autonomy and vice versa. Consequently, you cannot use AHP, because this method demands that both criteria must be independent

12 – “This was performed by identifying stability intervals for criteria weights”

This is really important because the strength of a solution depends of this stability. Now, how do you determine it? You don’t explain

13 – Page 11 “But in this case, since we have a lot of alternatives, the intervals of the weights will be very low. For example, if the weight of the criterion “Price” was changed to 70% instead of 38%, the model Dacia Spring Electric would be the most favoured electric vehicle since it is the model with the lowest price compared to the other models”

When you are doing sensitivity analysis (SA), you analyze each alternative individually, therefore, the number of alternatives is not related with intervals. When you perform SA, you have first to decide for each alternative which are the criteria that define it. Not all criteria intervene. If you have say, 9 criteria, it could be that only 2 are responsible for the best alternative, maybe 7 for the second best, 3 for the third best, and so on.

The criteria that in each case define an alternative are called ‘binding criteria’ and are the only important, you can ignore the others. Consequently, you are interested in determining the interval of weights only for the binding alternatives, not for the others.

The intervals of binding criteria do not depend of the alternatives, but on their lineal relationship to other criteria. In the example that you put, increasing the price criteria in a 70% from 38%, the model will select another model. First of all, do you know if the 70% is within the allowed interval of criterion price? Most probable, after a certain interval that criterion ceases to be binding, and then, it is irrelevant if you increase it even more its limit. Sorry, but your explanation is invalid.

14. In page 14 you speak about a walking weight, but you don’t explain what it is.

I guess that you are referring to a diagram as in Figure 10, where there are depicted straight lines representing alternatives, and how they increase or decrease their values when a single criterion weight, in this case consumption, is increased. In so doing, the consumption weight crosses all alternatives.

In your figure the larger the importance of the consumption weight the greater the advance of Tesla, because it is the car with the lowest consumption. Conversely, if the consumption weigh has a low importance, people is not concerned about it, and prefer the Jaguar. This is a trade-off between consumption and comfort.

However, this walking weight procedure does not represent reality, because only one criterion is chosen at random for variation and the others kept constant, and this is not real. This is the known ‘Ceteris Paribus” principle (Best between pairs) in Economics ,which most economists do not support. The reason is simple: The fact that consumption has the highest weight is not guarantee that it is the more important. Why?

Because there are other factors that condition that superiority, and thus, some criterion may be more important than the one with the maximum weight. True, people ranked criteria, but they do not consider the interrelationships; again, this is violation of system theory. The only way to know real importance is executing an entropy analysis, or using standard deviations.

In addition, it does not consider that, as long as consumption increases there are also changes in other criteria, that can modify the result. Who knows which is the interval for consumption?

I hope these comments can help you.

Nolberto Munier