Nos tutelles

PSL * CNRS Dauphine

Rechercher





Accueil > PRODUCTIONS SCIENTIFIQUES > Autres > Documents

UTA+ application (v 1.20) - User’s manual (M. KOSTKOWSKI, R. SLOWINSKI)

publié le

The UTA method was originally proposed by E.Jacquet-Lagrèze and J.Siskos in 1982. It has gained popularity thanks to implementation called PREFCALC, written by E. Jacquet-Lagrèze. Then, it was improved and modified by several authors.
UTA+ was written in Borland C++ 3.1 with Object Windows Library, at the Poznan University of Technology, Poland.
The UTA+ software is the latest implementation of the UTA method in Windows environment. It summarizes all important contributions made by other authors and offers some new possibilities, in particular, a "compensation" of marginal utility functions controlled by the user and the use of preference intensities in addition to the ranking defined by the user.
Inquiries about the program should be addressed to :
e-mail : vallee@lamsade.dauphine.fr
slowinski@pozn1v.put.poznan.pl
web : http://www.lamsade.dauphine.fr

General idea of the UTA+ method

The method can be used to solve the problems of multicriteria choice and ranking on a set A of alternatives. It constructs an additive utility function from a preference weak order defined by the user on a subset A’ of reference alternatives. The construction, based on a principle of ordinal regression, consists of solving a small LP problem. The software proposes marginal utility functions in piecewise linear form as compatible as possible with the given weak order. It allows the user to modify interactively the marginal utility functions within limits following from a sensitivity analysis of the ordinal regression problem. For these modifications, the user is helped by a friendly graphical interface.
The utility function accepted by the user serves then to define a weak order on the whole set A of alternatives.
The UTA+ software is composed of four main modules :
An example shows the intuitive ideas behind these developments.

    1.Problem editing, including definition of criteria and alternatives, and of the most and least preferred values of criteria.
    2. Definition of a ranking ( preference weak order ) on a small set of reference alternatives.
    3. Ordinal regression, including specification of required properties of the utility function prior to application of the solution procedure.
    4. Display and modification of the marginal utility functions and application of the accepted utility function for the calculation of the final ranking of alternatives.