A Stochastic Emergency Vehicle Redeployment Model for an Effective Response to Traffic Incidents

This paper studies the stochastic emergency vehicle redeployment problem for an effective response to traffic incidents. Both potential service demands and unavailable time of emergency vehicles already in service are treated under uncertainty. We develop a stochastic programming model for the problem, aiming at optimizing the system-wide performance by adjusting the scheduling plan to reposition emergency vehicles when some emergency vehicles become temporarily unavailable in response to service calls. An enhanced version of the L-shaped method is developed to solve the model. A new set of lower bound constraints are created to improve the quality of the lower bound. The computational results show that the proposed method yields a tighter lower bound and converges faster to the optimal solution than the conventional L-shaped method. A comparative analysis of different strategies in dealing with the unavailable times of busy emergency vehicles is conducted to assess the performance of the proposed model. The results indicate that better system performance can be achieved by explicitly incorporating the information about the status change of emergency vehicles currently in service into the redeployment plan.