摘要

Massive heterogeneous data processing has been a great challenge to intelligent traffic applications. In this paper, the dynamic shortest path problem in traffic guidance is dealt with, and a mathematic model of dynamic traffic networks is constructed. Then, a dynamic shortest path algorithm considering the intersection delay is proposed. Furthermore, a distributed and parallel processing model for solving the dynamic shortest path problem is presented based on HaLoop MapReduce and by using big data techniques. Finally, the proposed algorithm is tested on the intelligent traffic management and control platform based on continual flow. Experimental results demonstrate that the proposed algorithm and the presented model can effectively find the dynamic shortest path in large scale networks and can meet the real-time requirement.

全文