SOLVING MULTIOBJECTIVE DIET OPTIMIZATION PROBLEM UNDER UNCERTAINTY

The diet optimization problem considered in this paper concerns feeding farm animals. The nature of this problem requires taking into account multiple objectives on the one hand, and uncertainty following from imprecision of data on the other hand. Since the uncertainty has an epistemological character, it is natural to express imprecise data in terms of tolerance intervals with a most possible value (or subinterval) and decreasing possibility for other values within the interval. This corresponds exactly to the definition of fuzzy numbers. In consequence, the feeding problem is formulated as a multiobjective linear programming problem with fuzzy coefficients. In order to solve it, the interactive method FLIP proposed by Slowinski is applied. A large part of the paper is devoted to solving a real feeding problem with the FLIP software on a micro.