Collaborative Research: Designing Responsive Product and Service Fulfillment Networks using Dynamic Adaptive Modeling

该协作研究项目解决了一种新的大规模优化模型，称为具有服务保证的网络设计(NDSG)，该模型同时考虑了收入、成本和服务(端到端延迟)。虽然NDSG模型具有巨大的实际意义，但无论从理论还是计算角度来说，求解都是非常困难的。目标是使用最先进的优化技术来开发和测试有效的解决程序，这些技术可以利用NDSG问题的特殊数学结构。因此，该项目需要进行理论和经验调查，包括开发和描述替代问题公式，分析其结构特性，设计专门的解决算法，并使用实际数据测试这些方法。为了解决这一问题，将研究一族新的优化算法，该算法动态地重构问题，并相应地采用分解和模型强化技术相结合的求解方法。预计核心研究贡献--新的建模范式、新的解决方案方法和算法实现--也将扩展到其他大规模优化模型。在新的网络经济中，通信和物流能力变得至关重要。组织已经开始依赖这些能力来建立紧密集成的合作伙伴关系，从而有效地满足不同的市场对产品和服务的需求。信息或货物流动的延迟或失败不仅会对供应链造成破坏性后果，还会对紧急行动、空中交通管制和金融系统等服务造成不利影响。支持协调和协作所需的通信和分配网络的配置决策具有重要的战略意义，因为它们需要数十亿美元的巨额投资，而且网络的拓扑设计在很大程度上决定了它可以提供的服务水平。因此，优化工具在实践中被广泛用于设计产品和服务履行网络，而网络设计仍然是一个非常活跃的研究领域。设计最佳网络配置需要在各种相互冲突的目标之间进行复杂的权衡，例如最大化盈利能力、确保充分利用资源、提高服务级别等。经典的优化模型主要关注成本最小化，它们倾向于设计利用规模经济的稀疏网络，但在服务和可靠性方面并不健壮。该项目将为研究生提供进行论文研究的机会，并丰富优化和操作建模方面的研究生课程。

This collaborative research project addresses a new large-scale optimization model, called Network Design with Service Guarantees (NDSG), that simultaneously incorporates revenue, cost, and service (end-to-end delay) considerations. Though of immense practical significance, the NDSG model is very difficult to solve both from theoretical and computational perspectives. The goal is to develop and test effective solution procedures using state-of-the-art optimization techniques that can exploit the NDSG problem's special mathematical structure. Accordingly, the project entails both theoretical and empirical investigations, including developing and characterizing alternative problem formulations, analyzing their structural properties, designing specialized solution algorithms, and testing these methods using realistic data. To solve the problem, a new family of optimization algorithms that dynamically reformulates the problem, and correspondingly adapts the solution approach by combining decomposition and model strengthening techniques will be investigated. It is expected that the core research contributions-a new modeling paradigm, a novel solution approach, and algorithmic implementation-will also extend to other large-scale optimization models. Communication and physical distribution capabilities have become critical in the new networked economy. Organizations have come to rely on these capabilities to establish tightly integrated partnerships that can effectively serve diverse market needs for products and services. Delays or failures in the movement of information or goods can have debilitating consequences not only for supply chains but also for services such as emergency operations, air traffic control, and financial systems. Decisions regarding the configuration of communication and distribution networks needed to support coordination and collaboration have strategic importance both because they entail massive investments of billions of dollars and because a network's topological design largely determines the level of service it can provide. Consequently, optimization tools are widely used in practice to design product and service fulfillment networks, and network design continues to be a very active research area. Designing optimal network configurations entails complex tradeoffs between conflicting objectives such as maximizing profitability, ensuring adequate resource utilization, improving service levels, and so on. Classical optimization models, which focus primarily on cost minimization, tend to design sparse networks that exploit economies of scale but are not robust in terms of service and reliability. The project will provide opportunities for graduate students to conduct thesis research, and enrich graduate courses in optimization and operations modeling.