Hello Mr. Bajaj,

Your doubt is quite apparent as there are plenty of MCDM and MADM techniques in the literature. In fact, the most popular Multi-Attribute Decision-Making Techniques are WSM, AHP, ANP, TOPSIS, ELECTRE, GRA, PROMETHEE, ROC, MOORA, OCRA, to name a few. Though the fundamental logic of all the techniques is similar, the selection of the appropriate techniques is always important and thus demand justification. Generally, the choice of the right method is made based on the characteristics of the problem in hand. Since you haven't provided your problem description, no one will be able to suggest a particular tool. However, the following are some of those problem aspects that determine the appropriateness of the techniques:

1. If the problem on hand has sub-criteria, then choose AHP or ANP.

2. If the weights of criteria are dependent on alternatives (That is if a reverse relationship exists or if you need need to alter the weights assigned to criteria with respect to the alternatives), then select ANP instead of AHP.

3. In case you have Ratio/Interval/Ordinal scaled data corresponding to each alternative with respect to each criterion, (For instance, experimental results) then it is better to choose TOPSIS. Unlike most of the techniques like AHP, PROMETHEE, etc., TOPSIS is not based on pair-wise comparisons, making it most appropriate in those circumstances.

If it's not possible to assign a crisp relative score while trying pair-wise comparison (both between criteria and also between the alternatives), then you may deploy a fuzzy scale (Triangular fuzzy numbers/Trapezoidal fuzzy numbers) in order to transform the methodology more comprehensive.

For further understanding you may refer the following book by the famous Venkata Rao Sir (SVNIT Surat):

Rao, R. V. (2007). Decision making in the manufacturing environment: using graph theory and fuzzy multiple attribute decision making methods. Springer Science & Business Media.

I hope the information helps.

-Vishnu

Vishnu C Rajan Thanks for the answer.In my problem I have 17 drivers and 7 stages of product development.I need to identify which drivers can I assign to which stage and to assign these drivers for each stage depending upon their relevance for the stages.

Dear Saatyajeeet

As Vishnu said it is very difficult to help you if we don't know exactly which your problem is.

I am afraid that your last explanation is confusing. For instance what is it that you call 'drivers' and 'stages'

Recently, I published here in ResearchGate a guide for selecting methods, and which may help you. It is entitled:

300-Tool for selecting the most appropriate MCDM method to solve a problem

Nolberto Munier I have 17 principles(previously referred as drivers in my last answer) of lean manufacturing like 5S,poka yoke,kaizen,JIT,KAnban and so on.Now I need to assign these principles to 7 different stages of New Product development.Each stage may get assigned to more than one of these principles.So I was pondering about the MCDM which may be helpful in dealing with these problem.

Dear Saatyajeet

Did you notice that you change three times the name of an element ?

First there were alternatives, further you called them drivers and now principles. I assume that you are referring to alternatives, that can be principles. Is this correct?

Now, what are these strange names that you posted?

Are these the names of the alternatives?

What is it that you call ‘stages’?

Do you mean that each alternative can be assigned to more than one stage, whatever they are?

Or than the same alternative can be assigned to more than one stage?

You have to recognize that your problem is difficult to understand if you are not more explicit.

Where are the criteria?

Let me to tell you what I think your problems refer to. of course, if not, correct me

I assume that what you call ‘stages’ are in reality different scenarios and that the same alternative may be repeated in some of them. Is this correct?

I guess that the criteria explain the potential use of each alternative in each scenario. Is this so?

If this is the case, the problem is not so difficult to solve.

You can build a matrix with alternatives in columns and the seven scenarios in rows. Within each scenario you have to identify the relevant criteria, that is, criteria may be not the same neither in type nor in quantity.

Then, you have to assign values at the intersection of each alternative with each criterion considering that there could be no values for some alternatives and some criteria, just leave them blank.

You can use Promethee or Topsis to solve this problem, and you can get the criteria weights from entropy.

Or you can use SIMUS where no weights are necessary.

The problem could be more complicated if you have dependencies between alternatives such as inclusive or exclusive, , and then the only method that can handle this is SIMUS

I hope it helps, if not, let me know.

Nolberto Munier In my previous answer when I call it as "stages of new product development" then I'm referring to following:

1)Idea generation and idea screening

2)Concept testing

3)Business analysis

4)Product development

5)Test marketing

6)Commercialization

7)Review.

I have identified 17 lean principles(ways when employed gives a system which can be translated into lean) ;

some of them are - front load product development,5S,Establishing customer defined value,Defer elicitation,Kaizen,Value proposition,Leverage scope ,pareto principle,,Employee involvement and so on.

Now,I'm interested in finding a method which will help me in assigning these 17 principles to the above given stages (of new product development) provided none of the principle is assigned to more than one stage.

Hope I'm able to explain the problem.

Dear Saatyajeet,

Though the problem that you are investigating is clear from the last reply, I still doubt the criteria that you have identified to prioritize these listed 17 principles. Actually, you haven't provided any information relating to the identified criteria. Please understand that you need to identify a set of criteria (That too conflicting/contradicting in nature) in order to justify the application of any MCDM tool.

If you have those criteria, then the problem at hand is pretty simple. In fact you may follow the procedure described below:

Step 1: Prioritize the seven stages employing any of the MCDM tools

Step 2: Rank the principles (alternatives, 17 nos) corresponding to the stage with top priority (obtained from step 1).

Step 3: Remove the principle assigned to the stage (in step 2) and update the list of principles with the remaining 16 nos.

Step 4: Select the next stage and repeat step 2 and step 3 until all the seven stages are assigned with principles.

I very much understand that the above description is not the answer to your primary problem. For this, I strongly recommend you to go through the article and the Tool book (Excel sheet) published by Nolberto Munier mentioned in his first reply. This resource will definitely assist you in identifying the most appropriate MCDM tool for the above problem. It's very insightful and furthermore, the article is the most comprehensive one that one could get.

I wanted to thank Nolberto Munier Sir for this wonderful manuscript and tool book; it's really helpful.

Borrowing Sir's own words, "I sincerely hope that it can help."

-Vishnu

Dear Vishnu

I believe that your procedure regarding the problem submitted by Saatyajeed, may work, I don't know if this is possible and will solve his problem, but for me it makes sense.

I agree with you that our friend should clarify the problem a little more, although I believe that maybe I have a solution for that, using binary matrices.

I am thinking about that.

Thank you for your words, nothing will make me happier than helping people, and to tell you the truth, these questions and problems teach me a lot, because

illustrates problems and situations that I have not seen before.