News & Events


SIGSIM PADS 2018 Best Paper Award

Published on: 06-Jun-2018

From left: Dr Philipp Andelfinger and Prof Cai Wentong were receiving the award from Prof Francesco Quaglia (General Co-Chair), Prof Georgios K. Theodoropoulos (Program Chair), and Dr Alessandro Pellegrini (General Co-Chair).

Philipp Andelfinger, Yadong Xu, Wentong Cai, David Eckhoff, Alois Knoll. “Fast-Forwarding Agent States to Accelerate Microscopic Traffic Simulations”. 2018 ACM SIGSIM Conference on Principles of Advanced Discrete Simulation (PADS), May 23-25, 2018, Rome, Italy.

The above paper by researchers from the Parallel and Distributed Computing Lab at the School of Computer Science and Engineering at NTU, and the Area-Interlinking Design Analysis group in the TUMCREATE programme, received the award for best paper at the 2018 ACM SIGSIM Conference on Principles of Advanced Discrete Simulation (PADS). PADS is the flagship conference of ACM's Special Interest Group on Simulation and Modeling (SIGSIM). At NTU, the research was conducted by Research Fellow Dr Philipp Andelfinger and Prof Wentong Cai.

CityMoS – A City Mobility Simulator

Their paper is concerned with the high-performance execution of large-scale traffic simulations. In the TUMCREATE programme, researchers from fields such as traffic engineering, electrical engineering, and computer science collaborate to design a future transit system for Singapore based on electrical autonomous vehicles. To evaluate the effects of various design decisions on the overall traffic in Singapore, the simulation platform CityMoS was developed within the programme. CityMoS is able to simulate the traffic of Singapore at the granularity of individual vehicles.

In their award-winning paper, the authors propose a novel simulation time advancement method which avoids unnecessary iterative updates of vehicle positions and velocities throughout the course of the simulation. Instead of performing step-wise updates at each simulation time step, periods are identified during which simulated vehicles are isolated on a road segment. Using a "fast-forward" function derived from the most common simulation model of car-following behaviours, isolated vehicles are advanced in time in a single step, decreasing the computational cost of the simulation substantially. The authors show that the deviation from a traditional purely time-stepped execution is marginal. The proposed method enables more timely results in traffic simulation studies without impacting the fidelity of the results.

The research was funded by the Singapore National Research Foundation under its Campus for Research Excellence And Technological Enterprise (CREATE) programme.

Back to listing