Dear Colleagues
I am think of developing a rating model (somewhat like what the rating agencies use such as Moodys, S&P or Fitch) for an infrustructure sector in my country. This i reckon is a sorting problem. I have read somewhere about using absolute or rating scales in AHP as opposed to a relative (pairwise) modality, can this work? If not which is the best method for this sort of a problem?
Further, in an event that there are a number of project applications, i need this model to also be in a position to rank these projects (alternatives) with posibilities for some sensitivities. This i reckon is a ranking problem.
Which MCDA tool would be best to develop such a model. If possible include the online software suitable for any proposed MCDA tools.
AHP is a popular MCDA tool and widely accepted in practice. One can implement AHP with Excel or using software.
I concur with Ruhul - AHP is always a good first shot. Note that AHP is based on pairwise comparisons and not on absolute ranking mode. You may want to combine it further with ELECTRE or TOPSIS.
I suggest using the SIMUS method with which you can select and rank
I suggest using PROMETHEE method. Recently, this method is attracting many researchers and practitioners. It is easy to implement. Please have a look at this. The software provides visual interface and is free.
http://www.promethee-gaia.net/software.html
Dear Bruno, In fact AHP is a great and popular tool. Also is originally developed and mostly used with relative measures but there are a way to use AHP in absolute scales you can read about it in the original source Saaty, 2005. "Theory and Applications of the Analytic Network Process: Decision Making with Benefits, Opportunities, Costs, and Risks" RWS Publications. Pittsburgh (i.e. pages 20-23)
Hi Bruno,
Indeed AHP is a very famous method but based on your explanation, the first part of your project is a sorting problem.
I'm not sure if AHP is appropriate for sorting problems but it can certainly be used within a ranking context (mostly used for pairwise comparisons-based approaches ; although it also can be used with absolute scales as mentioned by Manuel).
There is AHPSort (sorting method based on AHP) but I'm not sure it's well suited since I have the impression that the profiles and the sorting context/parameters have to be very transparent and clear - i.e. in a rating system. Reference to AHPSort:
- https://www.researchgate.net/publication/238046864_AHPSort_an_AHP-based_method_for_sorting_problems?ev=prf_pub
Since I'm quite familiar with AHPSort, I would rather propose you to use the FlowSort method which allows you to use further Promethee in the second phase - FlowSort is based on the Promethee methodology. FlowSort is good for completely ordered categories. There is an extension if the data is not precisely defined (i.e. evaluations are defined by interval or fuzzy numbers).
Here some links:
- https://www.researchgate.net/publication/24065517_FlowSort_a_flow-based_sorting_method_with_limiting_or_central_profiles
- https://www.researchgate.net/publication/256606914_An_extension_of_the_FlowSort_sorting_method_to_deal_with_imprecision
FlowSort is implemented in the www.smart-picker.com software for which there is a trial version.
Hope this helps.
Philippe
Article AHPSort: an AHP-based method for sorting problems
Article FLOWSORT: A flow-based sorting method with limiting or centr...
Article An extension of the FlowSort sorting method to deal with imprecision
Phillippe thats it's what I call a very good and informative answer. Congratulations
Hello colleagues.
Thanks a million for the feedback, well appreciated. Like i indicated, i want to be very sure as i am putting together the methodology chapter for my PHD in infrustructure Financing and so far your advise is enriching.
- Badges
- Science topic