Mobility-As-A-Service for Resilience Delivery in Power Grids: Stochastic Programming Advancements under Decision-Dependent Uncertainties

Power grids are constantly subject to extreme emergencies that may leave communities without electricity and are critical threats to health and public safety. This NSF project aims to offer a new decision-making paradigm to deploy mobile power sources (e.g., mobile energy storage systems, mobile emergency generators, electric vehicles, and electric buses) for enhancing the resilience of communities and critical infrastructures when facing extreme events.

The project will bring transformative change to how uncertainties are managed when making allocation and dispatch decisions of mobile power resources (pre-positioning, routing, and scheduling) before and during disasters. This will be achieved by accounting for uncertainties that cannot be resolved simply with the passage of time, but whose resolution depends on whether or not the decision to deploy these flexible resources is actually made, i.e., decision-dependent uncertainty.

The intellectual merits of the project include devising new mathematical optimization models and expedient methods to solve the proposed problems promoting mobility-as-a-service for resilience in power grids. The broader impacts of the project include organizing an interdisciplinary seminar series on Advanced Stochastic Programming for Power Systems, undergraduate student involvement in the project research, dissemination of the project findings through publications, and research showcase to K-12 students in the Washington metropolitan area.

Going beyond the traditional service restoration problems with exogenous uncertainties, this project disrupts the current practice by developing new risk-neutral and risk-averse stochastic optimization problems and reformulations for service restoration that incorporate decision-dependent uncertainty (DDU). Due to the difficulty to solve such models—that typically take the form of stochastic mixed-integer nonlinear (polynomial) programming (MINLP) problems, the proposed research also focuses on the development of computationally efficient methods to solve DDU-embedded optimization models on mobility-as-a-service for resilience.

The proposed models will account for the coordination of mobile power sources with a variety of heterogeneous flexible resources in power distribution grids, network reconfiguration practices, repair crews, and transportation system constraints and schedules. The completion of this plan requires an interdisciplinary approach combining key methodologies from the mathematical programming and power systems engineering disciplines, thereby enabling convergence of multiple research areas.

This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.