GOALI: Global Meta-Hybrids for Supply Chain Optimization

这是一个与行业学术联络的资助机会（GOALI）奖。该项目的目标是开发全球元混合方法和相应的软件工具，以有效解决大规模供应链优化应用和其他大规模组合优化问题。这项研究工作是建立在最近对组合优化的启发式研究基础上的，包括全局元启发式嵌套分区（NP）框架。重点是问题类，可以为构建可行的解决方案和改进这些解决方案开发快速启发式。然而，与其孤立地考虑启发式，研究人员希望通过在基于分区的策略的全球框架内使用启发式，从启发式的可用性中获得最大的收益。全局元混合是指全局元启发式算法与线性规划技术的结合。这种混合方法以一种协同的方式将线性规划松弛的下界信息与启发式的上界和可行解数据结合起来。该项目将产生数学技术和相应的计算机软件，用于生成大规模供应链模型的高质量解决方案，以及由放射治疗等其他应用产生的大规模组合优化问题。这些问题对于目前的方法来说是难以解决的，因为它们没有利用在方法框架内利用的特殊属性。该项目是与罗克韦尔自动化公司合作的。

This is a Grant Opportunities for Academic Liaison with Industry (GOALI) award. The goal of the project is to develop global meta-hybrid approaches and corresponding software tools for the efficient solution of large-scale supply chain optimization applications and other massive combinatorial optimization problems. This research effort is built on recent research in heuristics for combinatorial optimization, including the global meta-heuristic Nested Partitions (NP) framework. The focus is on problem classes for which fast heuristics may be developed for both the construction of feasible solutions and for the improvement of such solutions. However, rather than considering heuristics in isolation, the researchers wish to obtain maximum benefit from their availability by employing them within the global framework of partition-based strategies. The term global meta-hybrid is used to denote the combination of a global meta-heuristic with linear programming technology. Such a hybrid combines the lower bound information available from linear programming relaxations with the upper bound and feasible solution data from the heuristic in a synergistic fashion.This project will yield mathematical techniques and corresponding computer software for the generation of high-quality solutions for large-scale supply chain models as well as large-scale combinatorial optimization problems arising from other applications such as radiotherapy. Such problems are intractable for current approaches that do not take advantage of the special properties to exploit within the methodological framework. This project is a collaboration with Rockwell Automation.