A DSS oriented method for multiobjective linear programming problems

Most of practical linear programming problems involve multiple and conflicting objectives. The paper presents an interactive method to approach this kind of problems. The main original aspect of this method lies in the fact that it combines the advantageous features of both the paradigms of Satisfactory Goals and Multiattribute Utility Assessment. It is a DSS oriented approach providing a ‘two level’ interaction: (1) interactive assessment of the decision maker’s utility function using the UTA ordinal regression model; (2) interactive modification of the satisfaction levels. Piecewise linear optimization techniques are used to determine, at each iteration, a new compromise solution over the set of efficient solutions.