Leveraging rapid simulation and analysis of large urban road systems on HPC. (February 2018)

• Record Type: Journal Article
• Authors: Turek, Wojciech; Siwik, Leszek; Byrski, Aleksander
• Is Part Of: Transportation research. Volume 87(2018)
• Page Start: 46
• Page End: 57
• Publication Date: 2018-02
• Subjects: Urban traffic simulation -- Parallel simulation -- High-performance computing -- Urban traffic control; Transportation -- Periodicals; Transportation -- Technological innovations -- Periodicals; 388.011
• DOI: 10.1016/j.trc.2017.12.014
• Languages: English
• ISSNs: 0968-090X

Abstract

Highlights: A transport-planning decision support system based on leveraging HPC for rapid simulation of the developments of the current traffic situation is presented. Scalable technology leads to efficient leveraging of the HPC appliances in order to get a very detailed prediction very quickly for huge simulation area. Tangible results of situation analysis consisting in detection and selected reaction strategies are described. The paper paves the way for further development of the system based on the verification of a seemingly already complete prototype system. Abstract: The scalable implementation of a microscopic simulation, presented in our previous work, opens new areas of applications for traffic simulation, namely short term traffic forecasting. It can be used for real-time prediction of local, exceptional situations. Moreover it can be used as a hypothesis verification tool for evaluating different strategies of traffic control. In order to realize such simulations in a very short time (what is crucial here) a sufficient computing power is necessary what can be achieved using dedicated HPC hardware. In this paper we present methods for rapid analysis of simulation results for large urban road systems, where a local, exceptional situation can remain unnoticed. We propose new metrics for detecting and locating such situations. The detected situation can be handled automatically with the method based on the multi-variant planning approach.