The global transition toward affordable, sustainable, and secure energy systems is deeply interconnected with the challenge of ensuring resilience in material supply chains. Significant portions of energy consumption stem from the production and transportation of energy-intensive materials, while rapidly growing sectors such as electric vehicles, renewable power, and AI-driven data centers, increasingly depend on critical raw materials. These dynamics create unprecedented stresses on both energy systems and supply chains, underscoring the need for unified, data-driven approaches capable of navigating uncertainty, optimizing resource allocation, and enhancing system resilience.
The overall research addresses these challenges by developing advanced, data-driven mathematical modeling frameworks that integrate predictive analytics with prescriptive decision-making under uncertainty. Despite advances in machine learning, big data analytics and statistical inference, a critical gap remains in translating predictive insights into optimization models that can guide operational, tactical, and strategic decisions in complex, uncertain environments. The proposed methodologies focus on quantifying uncertainties and embedding them into robust, flexible, and interpretable optimization frameworks applicable across both energy systems and material supply chains.
A significant component of this work contributes to the NSF-funded project “Securing Critical Material Supply Chains by Enabling Photovoltaic Circularity (SOLAR),” conducted in collaboration with the National Renewable Energy Laboratory (NREL). Our research focuses on developing mathematical models for reverse logistics & closed-loop networks in photovoltaic (PV) recycling, analyzing trade-offs between cost and emissions in end-of-life (EoL) management strategies [1,2]. We highlight the importance of system-wide, data-driven frameworks for PV recycling, upcycling, and reuse particularly for critical materials such as silver and aluminum by evaluating the economic and environmental benefits of various PV recycling pathways. In parallel, a decision-making framework and mathematical model have been developed to evaluate the impact of predictive analytics on supply chain performance, specifically assessing effects on total costs, inventory management, and demand fulfillment in photovoltaic (PV) recycling networks. This work illustrates how the integration of advanced data analytics, predictive modeling, and rigorous statistical validation can enhance supply chain resilience, cost efficiency, and sustainability outcomes [3].
Beyond supply chain applications, the research extends to energy systems planning and operation, coupling advanced forecasting techniques, such as Bayesian neural networks and probabilistic scenario generation, with stochastic and robust optimization techniques [4]. These models enable the generation of reliable, data-informed scenarios that support decision-making under variability, risk, and complex system constraints. Such integrated tools are critical for applications ranging from closed-loop supply chains and reverse logistics to long-term infrastructure planning, where balancing cost, environmental impact, and system resilience is essential.
Ultimately, this research delivers a unified, quantitative framework for designing and operating sustainable, resilient, and risk-aware systems spanning both energy and supply chain domains. By leveraging machine learning, statistical analysis, and mathematical optimization, it equips decision-makers and industrial partners with practical tools to transform complex data into actionable strategies, supporting sustainable innovation and resilience in the evolving landscape of global energy and material systems.
References