Optimizing ambulance dispatch and routing is among the most efficient ways for Emergency Medical Services (EMS) to save more lives at almost no extra cost. However, existing algorithms are designed for normal demand and often do not model hospital overflows. They are unable to adapt to disasters such as the COVID-19 pandemic, where case clusters emerge and hospitals rapidly reach capacity for weeks. Decisions optimized for normal times can suddenly become very inefficient, significantly delaying care. For example, in normal times, ambulances should take patients to the nearest hospital. However, in a pandemic, this strategy quickly inundates hospitals at the epicenters. A better system would use real-time data on incidents types and locations and hospitals’ capacities to optimally distribute patients across multiple equipped hospitals.

Professor Evdokia Nikolova and UT Austin College of Natural Sciences Professor Ngoc Tran’s Good Systems project, in collaboration with the Austin Travis County EMS Department, aims to create an optimal EMS dispatch system using real-time information. It leverages the latest data and models on COVID-19 to obtain the best available predictions on EMS incidents and hospital capacities, puts the problem in an optimization framework, and then solves it with risk minimization algorithms. The methods proposed would also solve the more general vehicle routing problem widely applied in supply chain logistics, which are similarly disrupted by the COVID-19 pandemic. By design, the proposed system can rapidly adapt to changing situations and is robust to disruptions. It minimizes risks to healthcare and the wider economy, preparing cities for future extreme events.