# 143

Dear Hamed Taherdoost

I have read your paper:

Analysis of Simple Additive Weighting Method (SAW) as a MultiAttribute Decision-Making Technique: A Step-by-Step Guide

My comments:

1- In page 1 you say “That is to say, the SAW as a classic version of the multi-attribute value method is “a value function is established based on a simple addition of scores that represent the goal achievement under each criterion, multiplied by the particular weights”.

In may opinion the underlined sentence is incorrect. The performance value in SAW represents how much a certain criterion contributes to an alternative, not an achievement

If you allow me, I propose an example:

Your company manufactures three products, A, B and C. One of the criteria establishes that the maximum capacity of production of your plant is, say for instance 200 units/month, in any mix of products. This is your target. It says that as a maximum you can fabricate 200 units, when yon take into account all the factors that intervene in the fabrication.

You run your MCDM method and it finds that your maximum production, considering your available monthly budget, wages, time, raw materials, quality control, rejects, etc., is 185 units. When you compare your maximum production of 200 units, to the real capacity of your plant of 185 units, you find an efficiency of 92%; this is your goal achievement. Obviously, you cannot make this rational analysis using SAW.

2- In page 2 “This method uses the idea of integrating the values of criteria and weights into a single value”

I guess that your sentence does not express what to want to say. Normalization, any method, converts all values to unitless values, normally between 0 and 1, not to a single value; it does not integrate anything

3. Sorry, but I do not concur with your applications for SAW. It is only a rudimentary method. However, it has the advantage that it is immune to rank reversal, and this is important. This immunity comes from the SAW structure that ranks alternatives independently, as it should be.

4- You mention as one advantage the fact that it allows compensation. In my opinion that is not realistic. For instance, can you compensate decreasing 0.2 points in a criterion like ‘environment’ to compensate a criterion like ‘disposable income’? You surely noticed that all the major MCDM methods do not use compensation.

‘Being intuitive’ an advantage? Well, probably in selecting a restaurant to dine or a movie, but not in a serious real-life scenario

These are my comments. I hope that they can be of help

Nolberto Munier