Implementation of MCDA methods in DecernsSDSS


Realization of MAUT and ProMAA methods is based on the original library for computation of functions of random variables (without implementation of Monte Carlo methods); and realization the 'fuzzy methods' within the DECERNS WebSDSS is based on the original library for computation of functions of fuzzy variables; these two libraries developed as a part of the DECERNS project.
When solving a specific multi-criteria problem, a DECERNS user has the opportunity to choose the appropriate MCDA method(s). If desired, the user may compare results after the implementation of several methods, including an analysis of uncertainties associated with the chosen MCDA approaches (Yatsalo, et al., 2007).
Several tools are used while implementing MCDA methods. These include value-tree development and editing for structuring multi-criteria problems, performance table creation and editing, value path and scatter plot analysis, and sensitivity analysis (sensitivity to weights, and to value function changing in MAVT). Uncertainties are addressed through weight sensitivity analysis (for MAVT, AHP, TOPSIS, PROMETHEE), value function sensitivity analysis (for MAVT, MAUT, ProMAA, FMAA, and F-MAVT), using random performance (for MAUT), random weights and random performance (ProMAA), and, accordingly, fuzzy weights and fuzzy performance (for Fuzzy-MAVT and FMAA).
A group decision support application (GDS) within the DECERNS WebSDSS allows the user(s) to create and process various types of surveys and questionnaires, voting mechanisms, while interacting with MCDA modules within spatial multi-criteria problem structuring; this tool will be added to WebSDSS and described later.