An inventory model with variable levels of quality attributes via geometric programming

In the paper an inventory model is proposed for a single product of a profit-maximizing producer where the level of a critical quality attribute of the product is a decision variable (e.g. the processing speed in megahertz for semiconductor chips). Two additional decision variables are the order quantity and the demand rate. The corresponding price as well as variable and fixed costs are assumed to be power functions of one or more of these decision variables. This assumption makes it possible to utilize geometric programming techniques effectively, and the global optimal solution is derived for the inventory model. In addition, extensive sensitivity analyses on primal and dual geometric programming problems for the inventory model are presented.