| dc.description.abstract | At present, police units are facing the problem of insufficient police force. This results in the poor connection of each duty, and increases the burden of basic level polices. Recent years, technology enforcement and smart transportation have emerged rapidly which could be used for criminal investigations and traffic observations. It includes three key technologies ＂communication module＂, ＂automation＂ and ＂charging system＂, which is also suitable for performing patrol duties. Therefore, this research plans to construct a patrol mode with mixed vehicles and UAVs to save police patrol time, achieve high-efficiency crime prevention, and deterrence effects.
This research makes use of the time-space network flow techniques to define the flow of police vehicle and UAVs, construct a vehicle-UAV mixed patrol route planning model. The research model is a large-scale multiple-commodity network problem with additional restrictions. Which belongs to NP-Hard problem. In terms of the solution method, the C++ programming language is used with the mathematical planning software CPLEX to solve the model. When facing large-scale problems in reality, it is difficult to use the mathematical planning software to find the best solution within a limited of time. For this reason, this research develops two Heuristic solutions-＂area cutting method＂ and ＂resource decomposition method＂. In order to evaluate the practicality of the model and algorithm, this study takes a police station to conduct a sample test. First, we analyze the sensitivity of different parameters. Next, we analyze the plan with scale and collocation. At last, we analyze the performance of the algorithm with different scales. The results show that this model can be used effectively, and the algorithm is very efficient. This indicates that the model can be used by police units as a reference for future smart patrol development when it is more mature. | en_US |