Two nonconvex minimization approaches for the problem of determining an economic ordering policy for jointly replenished items

This paper describes nonconvex minimization approaches for the problem of determining an economic ordering policy for jointly replenished items. After reducing the object function in the problem to a single-variable nonconvex function, two methods are proposed for solving the problem. One is the Successive Underestimation Method which utilizes effectively the properties of the function A/t+Bt. Another is a modification of the Relief Indicator Method. Different from other heuristic methods ever proposed for the problem, the methods developed in this paper can obtain a global approximate solution within any prescribed error bound as is liked. Results of computational tests show that both of the proposed methods perform equally well for many test problems.