RésuméWhen it comes to solve multi-criteria decision making problems, a number of techniques can be used. These techniques provide a global ranking of the alternatives considered in the decision problem. In a context where there is a high number of alternatives and where decision criteria are related to soft goals, the decision problem is particularly hard to solve, and this for two reasons. The first one is that techniques guiding the decision maker through the evaluation are generally based on pairwise comparisons of alternatives. When the number of alternatives is high, there is a high number of comparisons to perform, which therefore requires intensive work for the evaluation step. The other techniques that can be used without this problem require the evaluation of the alternatives to be performed beforehand, and the difficulty of evaluating to which extent soft goals are satisfied represents the second reason why the considered problems are hard to solve. This thesis provides an analysis of the use of artificial neural networks to improve the relevance of the rankings of alternatives delivered by traditional techniques aiming at solving the considered type of problems. These techniques as well as artificial neural networks are first approached from a theoretical perspective. Afterwards, a model using a combination of artificial neural networks and of the weighted sum model is built in order to recommend smartphones to particular consumers, with particular needs.
|Date de réussite||2017|
|Superviseur||Ivan JURETA (Promoteur), Monique Snoeck (Promoteur) & Anthony Simonofski (Copromoteur)|
Structuring and Solving MCDM Problems using Artificial Neural Networks
Amaral De Sousa, V. (Auteur). 2017
Thèse de l'étudiant: Master types › Master en ingénieur de gestion à finalité spécialisée en Analytics & Digital Business