Civil & Environmental Engineering Seminar Series
Event Details
Risk, Optimization, and AI: Quantifying Access to Essential Services
Abstract: Risk is central to how societies design infrastructure, allocate resources, and respond to disruption, yet it is often treated as a neutral measure of uncertainty rather than a fundamentally value-laden construct shaped by assumptions about what—and who—matters. This talk begins by framing risk as relational and explicitly normative, emphasizing that risk assessment is not only about probability and uncertainty, but also about how decisions encode priorities, trade-offs, and definitions of harm. I argue that these value judgments become highly visible when risk is operationalized through access to essential services. Building on this framing, I present ongoing work that uses data science and optimization to quantify access to essential services during disruptions. First, I introduce natural language processing methods that extract non-spatial dimensions of access—such as affordability, acceptability, and availability—from large-scale public text data, with applications to food access. Results show that conventional distance-based definitions of food deserts systematically underestimate access risk, and the methods provide a framework for jointly evaluating spatial and non-spatial dimensions of access. We provide evidence that food insecurity reflects a broader set of system design trade-offs—between travel burden, affordability, and service quality—rather than a small number of geographically isolated areas. The second half of the talk focuses on the restoration of electric power following large-scale outages, using optimization models to examine how different operational objectives encode value judgments and shape recovery trajectories. We perform sensitivity analyses to attribute observed inequalities in outage duration to three sources: outage location, restoration prioritization, and network topology. Our findings suggest that disparities in recovery are driven primarily by pre-existing exposure and system design, rather than by post-event decision-making.
Ben Rachunok is an Assistant Professor in the Department of Industrial and Systems Engineering at North Carolina State University, where he leads the Sustainable and Resilient Systems Lab. His research advances the theory and practice of risk analysis and resilience engineering, with a focus on how communities, infrastructure, and technologies interact under uncertainty. He develops computational frameworks that integrate optimization, simulation, and artificial intelligence to quantify access to essential services—such as electricity, mobility, communication, food, and water—before, during, and after disruptive events. His work spans foundational questions about how risk and resilience are defined and measured, as well as applied decision-support tools for infrastructure planning and disaster response. Dr. Rachunok holds a B.S. in Industrial and Systems Engineering from North Carolina State University, a Ph.D. in Industrial Engineering from Purdue University, and completed postdoctoral training in the Civil and Environmental Engineering Department at Stanford University.