Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

A GPU-accelerated Large-scale Simulator for Transportation System Optimization Benchmarking

Jun Zhang, Wenxuan Ao, Junbo Yan, Depeng Jin, Yong Li

TL;DR

This work proposes the first open-source GPU-accelerated large-scale microscopic simulator for transportation system simulation and optimization, and implements a set of microscopic and macroscopic controllable objects and metrics provided by Python API to support typical transportation system optimization scenarios.

Abstract

With the development of artificial intelligence techniques, transportation system optimization is evolving from traditional methods relying on expert experience to simulation and learning-based decision and optimization methods. Learning-based optimization methods require extensive interactions with highly realistic microscopic traffic simulators. However, existing microscopic traffic simulators are inefficient in large-scale scenarios and thus fail to support the adoption of these methods in large-scale transportation system optimization scenarios. In addition, the optimization scenarios supported by existing simulators are limited, mainly focusing on the traffic signal control. To address these challenges, we propose the first open-source GPU-accelerated large-scale microscopic simulator for transportation system simulation and optimization. The simulator can iterate at 84.09Hz, which achieves 88.92 times computational acceleration in the large-scale scenario with 2,464,950 vehicles compared to the best baseline CityFlow. Besides, it achieves a more realistic average road speeds simulated on real datasets by adopting the IDM model as the car-following model and the randomized MOBIL model as the lane-changing model. Based on it, we implement a set of microscopic and macroscopic controllable objects and metrics provided by Python API to support typical transportation system optimization scenarios. We choose five representative scenarios and benchmark classical rule-based algorithms, reinforcement learning algorithms, and black-box optimization algorithms in four cities. These experiments effectively demonstrate the usability of the simulator for large-scale traffic system optimization. The code of the simulator is available at https://github.com/tsinghua-fib-lab/moss. We build an open-registration web platform available at https://moss.fiblab.net to support no-code trials.

A GPU-accelerated Large-scale Simulator for Transportation System Optimization Benchmarking

TL;DR

This work proposes the first open-source GPU-accelerated large-scale microscopic simulator for transportation system simulation and optimization, and implements a set of microscopic and macroscopic controllable objects and metrics provided by Python API to support typical transportation system optimization scenarios.

Abstract

With the development of artificial intelligence techniques, transportation system optimization is evolving from traditional methods relying on expert experience to simulation and learning-based decision and optimization methods. Learning-based optimization methods require extensive interactions with highly realistic microscopic traffic simulators. However, existing microscopic traffic simulators are inefficient in large-scale scenarios and thus fail to support the adoption of these methods in large-scale transportation system optimization scenarios. In addition, the optimization scenarios supported by existing simulators are limited, mainly focusing on the traffic signal control. To address these challenges, we propose the first open-source GPU-accelerated large-scale microscopic simulator for transportation system simulation and optimization. The simulator can iterate at 84.09Hz, which achieves 88.92 times computational acceleration in the large-scale scenario with 2,464,950 vehicles compared to the best baseline CityFlow. Besides, it achieves a more realistic average road speeds simulated on real datasets by adopting the IDM model as the car-following model and the randomized MOBIL model as the lane-changing model. Based on it, we implement a set of microscopic and macroscopic controllable objects and metrics provided by Python API to support typical transportation system optimization scenarios. We choose five representative scenarios and benchmark classical rule-based algorithms, reinforcement learning algorithms, and black-box optimization algorithms in four cities. These experiments effectively demonstrate the usability of the simulator for large-scale traffic system optimization. The code of the simulator is available at https://github.com/tsinghua-fib-lab/moss. We build an open-registration web platform available at https://moss.fiblab.net to support no-code trials.
Paper Structure (32 sections, 4 equations, 13 figures, 27 tables)

This paper contains 32 sections, 4 equations, 13 figures, 27 tables.

Table of Contents

  1. Introduction
  2. Related Works
  3. Existing Simulators for Transportation System
  4. Existing Transportation System Optimization Methods
  5. The Simulator
  6. System Design
  7. Controllable Objects
  8. Metrics
  9. Transportation System Optimization Scenarios
  10. Experiments
  11. Conclusion
  12. Statement of Open-source Codes
  13. System Design, Implementation, and Comparison
  14. Preliminary
  15. Execution Process
  16. ...and 17 more sections

Figures (13)

  • Figure 1: (a) The framework of the proposed simulator. (b) The linked-list of the center line for vehicle sensing by one pointer operation. (best view in color)
  • Figure 2: The overview of the five transportation system optimization scenarios. (best view in color)
  • Figure 3: (a) The performance comparison with different sizes of road networks and number of vehicles. (b) Comparison of real-world and simulated average vehicle speeds. (best view in color)
  • Figure A1: An road network example used for simulation. (best view in color)
  • Figure A2: The Step() execution process of the proposed simulator. (best view in color)
  • ...and 8 more figures