Production Planning for Responsive Delivery Date Commitment

Abstract

As manufacturers strive to become more customer-oriented, they are reexamining their order negotiation systems and processes to ensure responsive but realistic delivery date commitments to customers. Several companies now use commercial Available-To-Promise (ATP) systems to integrate data across the sales and production planning functions of the firm to provide real-time information on product availability. However, successful supply chain performance requires not only data integration, but also proactive coordination of the activities across these functional areas. Motivated by a problem facing a commercial-pipe manufacturer, we present a model-based approach to integrate and systematically evaluate tradeoffs between order commitment and production scheduling decisions. We consider a manufacturer that produces and sells a wide variety of products to customers with different priority levels. Due to significant product changeover times, planners at the manufacturing firm decide on production campaigns for different product families several weeks in advance using planned inventory targets, and progressively assign customer orders to planned production as orders arrive. Residual production is inventoried to meet demand until the next campaign for that product. To support delivery date negotiations with customers, we develop an optimization model that concurrently considers order commitments and production plans. We formulate this production-delivery integration problem as a mixed integer program. Since, this problem is NP-hard, we develop a tailored solution approach that builds on tight lower bounds from LP-relaxations, based on several classes of valid inequalities, to determine strong upper bounds using a heuristic LP-rounding procedure. We evaluate and report the performance of the solution approach over a wide variety of random problem instances. On average, the solution algorithm yields upper bound solutions that are within 1.7% of the optimal solution.