Search

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2006.03.043 Byline: Elliot Bendoly (a), Doug Blocher (b), Kurt M. Bretthauer (b), M.A. Venkataramanan (b) Keywords: Internet; Inventory; Supply chain management; Optimization Abstract: Traditional 'Brick-and-Mortar' operations face the challenge of adapting to a new set of competitive rules made necessary by consumers who want the option of ordering electronically via the Internet. To satisfy these customers, firms must develop strategies that integrate their standard retail in-store channel with this relatively new on-line channel. Therefore, this research is designed to provide insights into supply chain inventory management strategies relevant to 'Clicks-and-Mortar' firms trying to satisfy both on-line and in-store sales. Specifically, this work considers the total cost implications of various inventory allocation strategies while maintaining target customer service levels. Analysis focuses on the development of models capable of handling new operating strategies made possible by electronic commerce. The implications of inventory risk pooling are considered in depth, revealing the existence of characteristics that determine whether completely centralized or decentralized policies are preferable. Author Affiliation: (a) Department of Decision and Information Analysis, Goizueta Business School, Emory University, Atlanta, GA 30322, United States (b) Department of Operations and Decision Technologies, Kelley School of Business, Indiana University, Bloomington, IN 47405, United States Article History: Received 10 November 2004; Accepted 14 March 2006

Nonlinear programming methods for multiperiod capacity planning in a manufacturing system comprised of a set of machines or workstations producing multiple products are introduced. The manufacturing facility is an open network of queues where capacity at each work station may be changed in each time period. Two nonlinear programming approaches for determining the timing and size of capacity changes are proposed. Under one model, total capacity costs are reduced so that plant congestion is controlled via upper limits on work-in-process. In another, a weighted sum of product lead times is minimized according to budget limits on capacity costs.

A branch and bound algorithm for planning capacity in manufacturing and computer networks is proposed. The algorithm globally minimizes a concave cost function over a single convex linear performance constraint and lower/upper bounds on discrete capacity variables. Reoptimization procedures are used to solve the subproblems more efficiently. Computational tests on problems with up to 100 variables indicate the efficacy of the algorithm for capacity planning.