GAQ-EBkSP: A DRL-Based Urban Traffic Dynamic Rerouting Framework Using Fog-Cloud Architecture

Details

• Creators:
  Du, Runjia ; Chen, Sikai ; Dong, Jiqian ; Ha, Pual (Youngjoun) ; Labi, Samuel
• Corporate Creators:
  Sky-Lab- AI & CAV Enabling Technologies ; Purdue University ; Center for Connected and Automated Transportation (CCAT)
• Subject/TRT Terms:
  • Urban transportation
  • Routes and routing
  • Architecture
  • Visual presentations
  • Traffic
  • Drivers
• Resource Type:
  Presentations
• Geographical Coverage:
  United States
• Corporate Publisher:
  University of Michigan. Center for Connected and Automated Transportation
• Description:
  DRL(GAQ)-EBkSP model based-on fog-cloud architecture is proposed to dynamically reroute the vehicles in large transportation networks graph attention mechanism to fuse information and extract relevant information to enlarge the learning efficiency Proposed model outperforms baseline model (without learning) under different scenarios with different RV numbers High efficiency of the RVs in the network is achieved by the trained model Future work Apply different priority standard to the framework Include different rerouting ratio (RV/BV) in the testing stage Use other similar RL-based models (LSTM-Q, GCQ) as the baseline models.
• Content Notes:
  Presentation at INFORMS Annual Meeting 2021, Anaheim, California
• Format:
  PPT
• Download URL:
  https://rosap.ntl.bts.gov/view/dot/66744/dot_66744_DS1.pdf
• File Type:
  [PDF - 3.13 MB]
• Collection(s):
  US Transportation Collection
• Main Document Checksum:
  urn:sha256:64b2b52113811b7f9bfcc1e293b616bf6e0b153bc20dde44a39e68b17fd5357e