Computational study of the relationships between feasible and efficient sets and an approximation

The computational difficulty of obtaining the efficient set in multi-object ive programming, specially in nonlinear problems, suggest the need of consi dering an approximation approach to this problem. In this paper, we provide the computational results of the relationships between an approximation to the efficient set and the feasible and efficient sets. Random problem gene ration is considered for different sizes of the feasible set and we study t he implications with respect to the number of objective functions and vario us kinds of objective functions. Computational experience with this approxi mation suggests that we obtain a substantial improvement when it increases the number of objective functions.