Performance assessment of vehicle interdiction strategies in a simulation based environment on a complex transportation network

Abstract

We consider the escape interdiction problem in a transportation network. In the absence of traffic in the network, the criminal/attacker tries to escape from the city using any of the shortest paths from the crime scene to any randomly chosen exit point. In the presence of traffic, the attacker chooses the optimal path, which takes minimum time to reach his destination. On the contrary, police/defenders try to interdict the criminal on his escape route. This is a challenging task for police with limited resources. Again, a real city road network is also complex in nature. First, we develop a simulation-based model for the optimal allocation of resources using the SUMO simulator. Next, we focus on developing a more advanced search strategy like routing with optimal resource allocation. We develop a novel meta-heuristic approach in a simulation environment to interdict the attacker in a dynamic crime scenario. Like the previous approach, the attacker follows the path with optimal time to escape from the city. In contrast, defenders try to catch the attacker regardless of the path which the attacker takes. The defenders aim is to maximize the interdiction probability. As time plays a vital role, we choose a meta-heuristic approach to provide quality solutions in a time-efficient manner. We test the developed meta-heuristic with a case study on the IIT Kharagpur map, India. We analyze the performance of the mentioned approaches using the SUMO simulator both in the presence of traffic and without traffic. We develop a linear regression model to generate optimal path in the presence of traffic. Here traffic is generated randomly in the whole network using the SUMO simulator.