| We consider a portion of the supply chain that involves stochastic production and inventory decisions. The production system is shared by several products and demands significant changeover times. For implementation purposes, we develop simple closed-form solutions (approximations) that can be installed in an existing ERP system. The first part of the dissertation focuses on minimizing work-in-progress (WIP) by minimizing cycle time. We construct efficient heuristics to set a near-optimal target utilization as well as to determine lot-sizes. Subsequently, we use a (Q,r) policy to manage system inventory consisting of finished goods and WIP. We study models for production-inventory systems before we develop heuristics to minimize system inventory cost subject to various constraints on service. The research is motivated by several real-world applications ranging from metal stamping to pharmaceuticals. |
