Best Practices: Modelling and Sensitivity Analysis in MCDM

Considering the questions posed by hundreds of practitioners in a scientific forum such as Research Gate (RG) over several years, it appears that there exists confusion about the different elements of a MCDM scenario, as well as the scope of them, and how to approach some real-life situations. This section tries to clarify them. Needless to say, it neither contemplates all possible problems nor clarifies all doubts, but gives solutions to which are considered the most common and frequent qualms formulated by students, practitioners, and professors in this area. It shows, using numerous examples, how to model in MCDM real problems into the initial decision-matrix, and thus, replicating the scenario under study. In so doing, it fills a void in the published literature where it is noticeable that the scenario is not replicated in its whole dimension, which may lead to arguable results. The chapter tries to give answers to some questions, and mainly to point out certain features which are common to most projects and how to incorporate them in the modelling. No formulas are used. Instead, there are real-life examples, common sense, and reasoning, portraying actual scenarios, reflecting real issues, and solved either by Linear Programming (LP) or by SIMUS, a multi-criteria hybrid method, which can deal with most real features. These methods are not perfect, far from it, and perhaps there are better MCDM approaches, however, so far, to his author knowledge, no other method exits able to solve these problems, other than the two mentioned.