I am struggling to find a clear source of information on-line that will help me understand how to convert a fuzzy weight for a linguistic preference to a crisp value.
For instance, below we have a fuzzy scale that transforms linguistic preferences into numerical values on a scale [1...10].
Six levels are introduced with the linguistic scale: "Negligible (N)", "Very Low (VL)", "Low (L)", "Average (A)", "High (H)", "Very High (VH)".
Each of these levels represents the level of importance that is assigned to criteria in a multi-criteria decision making (MCDM) model.
Say for instance, we have one criteria that is given a VH importance. In this case, how does one determine the crisp value for the VH fuzzy weight? Is it just the centroid value of 9? Does the correct calculation involve using the triangular fuzzy number (8,9,10) for this example.
If you can confirm or point me to a good reference, I would be most grateful.
