The EMS vehicle patient transportation problem during a demand surge

We consider a real-time emergency medical service (EMS) vehicle patient transportation problem in which vehicles are assigned to patients so they can be transported to hospitals during an emergency. The objective is to minimize the total travel time of all vehicles while satisfying two types of time window constraints. The first requires each EMS vehicle to arrive at a patient's location within a specified time window. The second requires the vehicle to arrive at the designated hospital within another time window. We allow an EMS vehicle to serve up to two patients instead of just one. The problem is shown to be NP-complete. We, therefore, develop a simulated annealing (SA) heuristic for efficient solution in real-time. A column generation algorithm is developed for determining a tight lower bound. Numerical results show that the proposed SA heuristic provides high-quality solutions in much less CPU time, when compared to the general-purpose solver. Therefore, it is suitable for implementation in a real-time decision support system, which is available via a web portal (www.rtdss.org).