IZAR - multicriteria decision support system

Many of real decision making problems are evaluated by multiple criteria. For applying of appropriate multicriteria methods it is necessary a software support. The paper presents a universal multicriteria decision support system IZAR. The main component of the system IZAR is an expert system helping the user with choosing the method most suitable for available information about the problem. IZAR not only suggests the most suitable method but also applies it immediately to the problem solving. By means of properly specified questions the expert system selects the right procedure for solving the problem, when it takes into account its peculiarities regarding entries and additional information, which user can assign to the system. An appropriate classification of multicriteria models and methods is needed because of universality of the system. The system solves discrete and continuous problems. Basic models for multiattribute evaluation and multiobjective optimization problems are included. Methods for multiattribute evaluation problems are classified by types of preference information and by calculation principles. Preference information is given as aspiration levels, ordinal information, and cardinal information. Basic calculation principles are utility maximization, minimization of a distance from the ideal alternative, and evaluation by preference relation. Methods for multiobjective optimization problems are classified by means of setting user information, as a priori information, a posteriori information, and progressive information. This project is supported by FRVS - grant no. 2949/2006. The system will be available on web pages for all interested users and it will be also distributed on CD for users without internet connection.