A multilevel methodology for supply chain network design

点击翻译按钮获取中文摘要

This research programs aims to contribute to the effort of designing and planning more efficient, resilient and sustainable supply chains. The focus of this proposal is on developing advanced tools and planning systems to support and enhance strategic planning in supply chains. Methodologies supporting this research are based on optimization, distributed decision making, and multi-agent systems. Designing a supply chain (SCND) involves many interrelated decisions, such as selecting the number, location, mission assignment, capacities and technologies of its production-distribution facilities, selecting the key suppliers, and often determining which of the many potential product-markets the company should serve and what offer should it make to that market. This problem has significant challenges, but also provides significant benefits: in many industrial contexts, a better supply chain structure leads to lower logistics costs, better customer service and shorter lead times, allowing the company to increase its market share and profitability. Independent studies have estimated the total reduction of supply chain costs from redesigning a supply chain to be in the 5 to 15% range.  SCND involves two main activities: generation of candidate designs and the evaluation of those designs. More recent surveys show that despite the vast literature on the topic, there is very limited work on design evaluation beyond scenario planning. The proposed research will develop new methodologies to estimate the real-world performance of supply chain designs under varying operational contexts ranging from normal conditions to severe disruptions. This will in turn allow to improve the design process itself.