Dear Yonas

In reality any MCDM methods, except , in my opinion, AHP and ANP, may handle this problem efficiently. In reality it is simple

Dear Nolberto Munier

Thank you very much!

Do you think by using hybrid technique of AHP and ANP could lead to optimum alternative (Such as FAHP, DANP)?

Thank you in advance

I wonder how you can talk about optimums using AHP and ANP, when both are based on invented weights and umproved assumptions, like the concepts of impacts and feedback, mentioned by Saaty, but never explained.

As a matter of fact there is no MCDM that produces optimal values, because it is impossible to maximize benefits and, at the same time, minimize costs. And this is valid for all MCDM methods, and of course, applicable to FAHP. By the way, did you know that Saaty, the creator of AHP, said that you can't use fuzzy in AHP because it s already fuzzy?

Regarding DANP, that is using DEMATEL with ANP, is not a guarantee of best solution, because DEMATEL only determines causal relationshipons between criteria.

The selection of tractors depends on many factors that must be met in the field, for example the type of soil, the type of crop, the altitude, the slope of the terrain; as well as the weather conditions.

Dear Fredy

Of course, and they are not a matter of decisions, except perhaps the type of crop.

What role do AHP and ANP play here with their absurd use of pair-wise comparions and invented preferences?

A farmer, based on his experince may have very good arguments regarding the relative importance about those factors, nobody can deny that, but it does not mean that he can say for instance that slope is 3 times more important than altitude. This would be illogical and irrational

He can say for instance that according to his experience and consulting other farmers, that tractor XX is better than tractor yy in certain kind of terrain, and this is very valuable, but this is a post-result process, once the mathematical result using a MCDM method is known, and even he can dismiss the selection done by the mathematical model. But what he can't do is to change arbitrarily the exsting and real values at the sgtarting of the process as AHP qnd ANP do, and assuming that what is in his mind is applicable tothe real world

Dear Darius

Regarding your assertion, I strongly diaagree.

AHP is only relatively valuable for simple projects.It was not built for even semi complex problems. An example is that it can't work with interrelated criteria, as you have in selecting farm machinery, albeit ANP, from this point of view, can do the job. Another reason is AHP linear hierarchy, that is inexistent in nowadays scenarios

In my opinion, other than in personal projects where the consequence of the result will fall on the DM, AHP is useless, and in certain extent also ANP, because both are using some very controversial techniques as is pair-wise comparisons, which does not have any mathematical foundation and depend of the feelings and moods of the DM.

There is no room for preferences in real-life problems, you must work with real data, objective and subjective.

Where it is said that a DM can vote for thousdands or millions of people?

As a matter of fact there is a theoreme, The Arrow's Impossibility Theoreme that concluds exactly the opposite, that is, you cannot vote for others, which is what AHP and ANP are proposing. You do not need a theoreme for that, just use common sense.

Would you like that somebody that you did not choose, meet or know, vote in your place on a matter that is of your interest?

Please, think about it

Dear Prof. Nolberto Munier

To analyze both qualitative and quantitative data TOPSIS or HYBRID TOPSIS methods are preferred in this aspect. What is your expert opinion?

I agree, depending on how qualitative data is collected

At this respect , and in addition to the techncical, econmic and en ironmrntal criteria, I would perform a survey among the farmers, just asking them, in a 1 to 5 scale, to value each criterion independently, for instance reliability, easiness to get spares, most common failures, etc, and then, finding an averagee, but more important, trying to take advantage of their experience. Of course, I am referring to the performance values that express the contribution of each type of tractor to each criterion, not the value of the criterion itself, for which I would use entropy.

Dear Nolberto

You are right.

Thanks for your opinion.

You are most welcome Darius