Supply chain optimization under uncertainty

供应链规划系统中的许多挑战来自几个不确定因素的存在，例如需求、交货时间、服务和运营成本等。如果不考虑这些不确定性，可能会导致服务水平下降、需求减少以及从计划外的结果中恢复所需的大量额外费用。我们的研究重点是供应链系统的决策算法设计，这些算法可以从历史数据中提取不确定性，并可以利用数据分析来帮助规划者做出风险明智的决策，并主动对不确定性做出反应。第一个研究主题侧重于分销网络设计和库存优化的集成，这是供应链战略和战术规划中的两个重要组成部分，它们具有很强的相互作用。本研究主题涉及不确定性下优化的多种解决方法，以及数据挖掘技术的增强。第二个研究主题是关于随机组合优化的求解算法的发展，这是运筹学中一个非常具有挑战性的问题，可以应用于几个物流和运输应用，如车辆路线。第三个研究主题涉及动态生产和库存路径的算法设计，可用于处理需求等重要输入以动态方式显示的情况，以及操作环境或输入立即发生变化的情况，因此必须快速确定修改后的解决方案。我们的目标是1)提供有效的跨学科解决方案框架来处理这些具有挑战性的问题，2)同时专注于实际和创新的应用，这些应用可以导致该领域的新研究发展，3)为供应链从业者提供跨不同解决方案方法的见解和含义。我们计划在一些项目上与工业伙伴合作，他们将为科学发展提供测试实例和实用观点。

Many of the challenges in supply chain planning systems stem from the presence of several sources of uncertainty such as demand, lead time and costs of services and operations to name a few. Failure to account for these uncertainties can result in a reduced service level, loss of demand and significant extra expenses to recover from unplanned outcomes. Our research focuses on the design of decision algorithms for supply chain systems that can incorporate uncertainty which can be extracted from historical data and can make use of data analytics in order to help the planners make risk informed decisions and proactively react to the uncertainty. The first research theme focuses on the integration of distribution network design and inventory optimization, two critical components in the supply chain strategic and tactical planning that have strong interactions. This research theme involves multiple solution methodologies for optimization under uncertainty together with enhancements from data mining techniques. The second research theme concerns the development of solution algorithms for stochastic combinatorial optimization, a very challenging problem in operations research, which can be applied to several logistics and transportation applications such as vehicle routing. The third research theme involves algorithmic designs for the dynamic production and inventory routing that can be used to deal with situations where important inputs such as demand is revealed in a dynamic fashion, as well as when there are immediate changes in the operational circumstances or inputs, and thus a revised solution must be quickly determined. We aim to 1) provide efficient interdisciplinary solution frameworks to deal with these challenging issues, 2) while focusing on practical and innovative applications that can lead to new research developments in the fields and yet 3) providing insights and implications across different solution methodologies for supply chain practitioners. We plan to work on some of the projects with industrial partners who will provide test instances and practical points of view for the scientific developments.