查看文章

An optimal value bounds procedure is briefly described. A nonconvex program is considered. Only nonconvex problem functions need to be separable. The parameters perturbed appear in the constraints as well as in the objective function. The unperturbed problem is solved by a branch and bound algorithm. The convexity of the underestimating problems defined by the algorithm allows the construction of lower bounds on the optimal value function of the original problem. Two methods for calculating optimal value upper bounds are presented. An application problem on the determination of a product mix model when learning effects are present is solved.