Application of fuzzy nonlinear goal programming to a refinery model

Abstract A simplified oil refinery model has been formulated as a fuzzy nonlinear goal programming problem, in which four non-compatible performance criteria (objective functions or goals) exist beside ten crisp constraints in the form of material balance equations. Total yearly profit of the refinery and the sensitivities of the profit to variations in refinery conditions have been assumed to be fuzzy goals. Linear and S-type membership functions have been assumed separately for all the fuzzy goals. Box's complex method has been used to solve the crisp equivalent of the fuzzy nonlinear goal programming problem. The “min” operator has been used as the aggregator. A software developed in C implements the model. The results show that the present methodology gives the decision maker a good flexibility in setting up the goals, in that he/she is not forced to specify goals crisply simply for mathematical reasons. Also, the present treatment of the problem in a fuzzy framework enables the decision maker to consider any number of goals in any combination.