What do you mean by "criteria are incomparable"? on the ELECTRE Family, what we have is Incomparalbe Alternatives.

I believe all outranking methods (European School mainly), use this kind of technique.

If the data are numerical, you can use TOPSIS (and a many extensions of this method). If the data are qualitative (or "numerical and qualitative") you can use AHP or ANP methods. I will recommend a more adequately MCDM method if you expand the description of your problem.

"incomparable criteria" means that DM can not or don't want to give any information about criteria... there are no weight of the criteria.

AHP ANP TOPSIS methods need this information.

Yes, TOPSIS, AHP and ANP methods have need these information, but I didn't expect such an interpretation of your problem. If decision-maker don't give you any information, what you know about problem? Do you want obtain in result: ranking, the best alternative or clustering?

I think about methods which let support DM at incomparability of the criteria. Methods like ELECTRE IV, MINMAX, etc. It is here not important to me what area of problems (best solution, ranking, clustering or classification) are taken under consideration.

So, do you mean weightless methods?

Is the stochastic method possible? You cant inform a precise number, but you can try a probability to them.

Depending about the problem, you can assign the same weight for all criteria, and that means weight is not possible.

Here you can found another version of ELECTRE IV, called ELECTRE IV-h.

http://dx.doi.org/10.1590/S0101-74382007000200004

By the way, do you know the SMARTS Method? It is a kind of swings and exchanges of a compesantion and addictive procedure.

Maybe here you can find some direction...

Edwards, W. e Barron, F. H. (1994), SMARTS and SMARTER: Improved Simple Methods for Multiattribute Utility Measurement, Organizational Behavior and Human Decision Processes, 60, 306-325.

Thank you for responses.

What about vikor method?

I think it is usefull for incompatible criterias.

The vikor method needs importance weights, so, it isn't a example of weightless method. Please look for example section 3.2 in the paper:

Hu-Chen Liu, Long Liu, Nan Liu, Ling-Xiang Mao. 2012. Risk evaluation in failure mode and effects analysis with extended VIKOR method under fuzzy environment, Expert Systems with Applications 39(17): 12926-12934.

you are right, the vikor method needs Weights,

what about using same wrights ?

Maceij

You don't need to create weights for criteria. You can use Shannon's entropy method that gives exact weights based on  information contained in the problem, that is, without subjectivity. Zavadskas has an example of this