Optimizing the location of incident response vehicles for congestion mitigation

In this study, we study the location routing problem for determining the waiting locations of incident response vehicles, i.e. tow trucks. Incidents that occur at different points in a city, require tow trucks and cause traffic congestion at different volumes in the regions where they occur. How different incidents may affect the traffic can be determined by the sensors placed on the roads. Different number of incidents, different occurrence times and locations of incidents and different sizes of that traffic congestion under they cause are handled through different scenarios under the scope of this thesis. Considering different incident scenarios, the locations where the tow trucks should wait during the day are aimed to be determined in a way that minimizes the intervention time from the occurrence time of incident to the dispatch time and the traffic congestion caused by the incidents. To address the problem, first a mixed integer mathematical programming model has been developed and validated. However, because of that quality solutions could not be provided in a reasonable time by mathematical model, heuristic solution approach Greedy Randomized Adaptive Search Procedure has been developed. Incident scenarios have been created by use of real data obtained from Istanbul Metropolitan Municipality and realistic data sets have been created by analyzing the congestion spreading effect of the incidents. We demonstrate the performance of the proposed heuristic algorithm and its two different versions through realistic problem instances.