Abstract The COVID-19 pandemic resulted in significant supply chain disruptions across many industries, with disruptions caused by both increases in demand and reductions in available supply, leading to detrimental effects that disproportionately affected vulnerable populations. Further complicating matters, the severity of the disruptions varied throughout the pandemic. It was unclear how well data-driven models could capture the variations in time and at what point(s) in time different risk mitigation actions or operational decisions would be most effective. This project examines the effect of variations in risk over time on the management of transportation and supply chain networks with the goal of creating enhanced decision support tools. In the first year, the project addresses two related tasks: (1) to develop a set of models to represent the risk factors and time elements pertinent to common supply chain risk mitigation strategies, and (2) to examine the tradeoffs between prediction accuracy and prediction lead time.
Description "The tasks are listed within the research questions of the project.
(1) Models mapping risk factors and time elements.
(1.a.) Literature review 1: Conduct a literature review to include topics of supply transportation and supply chain network risk mitigation actions/strategies and decision models, dynamic risk management, and value functions. The literature review of risk mitigation strategies and existing decision models is expected to comprise the majority of the focus of this task.
(1.b.) Cataloging decisions: The results of the literature review will be used to create a catalog of risk mitigation decisions and common operational decisions with impacts on risk strategy. Because not all decisions may be conducive to analytical representation within a value function, the appropriateness of each decision for the development of an associated value function will be evaluated. For those decisions that are more conducive to analytical representation, the relevant parameters will be cataloged with each decision, including parameters to describe the potential outcome(s) of each decision.
(1.c) Model development: For each risk mitigation or operational decision identified as being more conducive to analytical representation, a value function will be developed or the situation will be mapped to an existing value function in the literature. For example, increasing inventory is an action that can be taken to mitigate supply disruptions, and many representations of this decision exist in the literature (Qin et al. 2011; Canyakmaz et al. 2022). The result will be a set of time sensitive value functions to represent common decisions in supply chain risk mitigation.
(1.d) Write-up of work 1: The results of tasks (1.a.) through (1.c.) will be presented in a white paper.
(2) Tradeoffs between prediction performance and prediction lead time.
(2.a.) Literature review 2: This portion of the project will include a literature review focusing on issues of changes over time within the context of predictive models and algorithms and will focus on issues such as concept drift.
(2.b.) Simulation development & sensitivity analysis: A Monte Carlo simulation will be developed to facilitate the characterization of tradeoffs between prediction accuracy at different points in time and value to the decision maker. The simulation will be used to conduct sensitivity analysis on relevant parameters. (2.c.) Analysis of results and write-up of work 2: The results of the simulation and sensitivity analysis will be characterized. Results of tasks (2.a.) and (2.b.) will be presented in a manuscript that will form the basis of a peer-reviewed journal article."
Objective "The long-term goal of this project is the development of dynamic decision support tools that adapt to shifts in observed risk factors and that are tailored to the risks common within transportation and supply chain networks.
Within the first year of the project, the expected results include the development of decision models that enable the analysis of time-varying risks posed to transportation and supply networks, including the analysis of changes in the effectiveness of risk mitigation strategies. These models will provide a foundation for on-going, future work in the development and refinement of dynamic decision support tools. The deliverables in the first year include a manuscript describing the literature review, decision model(s), and simulation results. The manuscript will be submitted to an academic journal. Target journals include the European Journal of Operational Research and Risk Analysis: An International Journal."
Impacts/Benefits "As evidenced by the economic turmoil that followed the disruption of transportation and supply chain networks during the COVID-19 pandemic, the ability of U.S. supply chains to respond to emergent risks affects the economic security of the country. The development of dynamic decision support tools to enhance the ability to respond to emergent risks and time-varying risks will reduce the negative impact of such disruptions. Ultimately, such advancements promote greater economic security and global competitiveness of US companies.
This project also promotes equity. Supply disruptions during the COVID-19 pandemic disproportionately negatively impacted vulnerable populations. By supporting the development of tools to promote transportation and supply chain resiliency, these disproportionately negative impacts will be lessened, thus promoting greater societal equity."