(15g) Towards a Multi-Stage Stochastic Optimization Approach for Resilient Supply Chain Network Design and Operations

As supply chains extend their reach and complexity, they become increasingly vulnerable to various risks and disruptions, such as extreme weather events, natural disasters, cyberattacks, geopolitical conflicts, and infectious diseases [1, 2]. Events like the COVID-19 pandemic and other low probability-high impact disruptions have underscored the weaknesses in the infrastructure and management of global supply chains, revealing their inability to effectively respond to such unprecedented risks [2]. Moreover, predicting the occurrence and consequences of such "black swan" events is extremely challenging [3, 4]. To effectively tackle the challenges arising from the heightened frequency and intensity of disruptions, it is essential to design and operate the next generation of supply chains while hedging against the uncertainties arising from both endogenous and exogenous factors.

To address this, we propose a multi-stage stochastic optimization approach that accounts for disruptions and their discrete probability of occurrence. To allow for time-dependent uncertainty realization, the planning horizon is divided into multiple time periods, each containing uncertainty realizations corresponding to the current stage. This further enables the simultaneous consideration of uncertainties at the tactical, strategic, and operational decision-making levels. The optimization formulation captures trade-offs between economic factors (first-stage costs and expected subsequent-stage costs) and resilience objectives (such as service level) through a multi-objective approach and is illustrated by a case study involving a multi-echelon distribution network [5]. The case study is setup in a python programming environment using the energiapy [6], pyomo [7], and mpi-sppy [8] packages. The results elucidate the dynamic reconfigurations of the supply chain network in anticipation of and during supply chain disruptions.

Keywords: Supply Chain Network Design, Uncertainty, Stochastic Optimization, Resilience

References

[1]. El-Halwagi, M. M., Sengupta, D., Pistikopoulos, E. N., Sammons, J., Eljack, F., & Kazi, M.-K. (2020). Disaster-Resilient Design of Manufacturing Facilities Through Process Integration: Principal Strategies, Perspectives, and Research Challenges. Frontiers in Sustainability, 1.

[2]. Iakovou, E., & White, C. (2020). How to build more secure, resilient, next-gen US supply chains. Brookings Institute TechStream; https://www.brookings.edu/techstream/how-to-build-more-secure-resilient-next-gen-u-s-supply-chains/ .

[3]. Gopal, C., Tyndall, G., Partsch, W., & Iakovou, E. (2023). Breakthrough Supply Chains: How Companies and Nations Can Thrive and Prosper in an Uncertain World. McGraw Hill Professional.

[4]. Bechtsis, D., Tsolakis N., Iakovou E., Vlachos D., 2022. “Data-Driven Secure, Resilient and Sustainable Supply Chains: Gaps, Opportunities, and a New Generalised Data Sharing and Data Monetisation Framework”; International Journal of Production Research; Vol. 60, No. 14, pp. 4397-4417;

[5]. Ivanov, D., Pavlov, A., & Sokolov, B. (2014). Optimal distribution (re)planning in a centralized multi-stage supply network under conditions of the ripple effect and structure dynamics. European Journal of Operational Research, 237(2), 758–770.

[6]. Kakodkar, R., & Pistikopoulos, E. (2023). Energiapy-an Open Source Python Package for Multiscale Modeling & Optimization of Energy Systems. In 2023 AIChE Annual Meeting. AIChE.

[7]. Hart, William E., Jean-Paul Watson, and David L. Woodruff. "Pyomo: modeling and solving mathematical programs in Python." Mathematical Programming Computation 3(3) (2011): 219-260.

[8]. Knueven, B., Mildebrath, D., Muir, C., Siirola, J. D., Watson, J. P., & Woodruff, D. L. (2023). A parallel hub-and-spoke system for large-scale scenario-based optimization under uncertainty. Mathematical Programming Computation, 15(4), 591-619.