Αναλυτικές μέθοδοι και πολυκριτήρια συστήματα υποστήριξης αποφάσεων υπό αβεβαιότητα : το σύστημα Τάλως
Χριστοδουλάκης, Νικόλαος Γ.
KeywordsΛήψη αποφάσεων ; Συστήματα υποστήριξης αποφάσεων ; Πολυκριτήρια ανάλυση ; Μεταβελτιστοποίηση ; Γραμμικός προγραμματισμός
This Decision Support System is designed to help the decision-maker in situations where he wishes to rank several actions (alternatives, proposals, strategies, projects, etc.) defined on several criteria. The performance of these actions on any criterion can be known with a certain level of uncertainty in the sense that they are characterised by a probability distribution on the variation interval of the criterion. In general terms, the DSS provides a ranking of all the actions which are assessed with a so-called ‘utility’ function built to represent the decision-maker’s system of preferences. This DSS is programmed with Visual C# on the Microsoft Visual Studio .NET 2012. We decide to program the application with the Visual C# because this language is the most applicable to satisfy all of our requests. We needed a programming language that could do fast a large number of mathematical operations, with less lines of code. C# is a structured language, which allows to spit big programs to small classes of code. It has a lot of the characteristics of a high level language, but can also use the same programming details, like a symbolic language. Also, C# accomplished all of our requests of the part of graphics that needs a GUI application. To solve linear programs resulting from the application of methods Stochastic UTA, SMAA, UTA GMS and Extreme Ranking, we created a LP Solver which takes as input a Simplex table, and then solve it. The Simplex Solver we created also used during post-optimal analysis. The application has 50 forms and 9 external classes. To total number of the lines of code for the whole application is about 20.000.