Multiobjective possibilistic linear programming

This paper presents two solutions to multiobjective linear programs when all the parameters may be fuzzy variables whose values are restricted by a possibility distribution. In the first solution, the possibility distribution of the objective function is first found; and then a solution consists of those values of the decision variables which produce the largest (undominated) values of the objective function, where these values also have maximum possibility. In the second solution, the possibility distribution of the undominated sets is first found; and then a solution consists of those values of the decision variables, which are undominated with maximum possibility, but also produce the largest (undominated) values of the objective function. For certain multiobjective possibilistic linear programs, it is shown that the two solution concepts are identical.