Agent-Based Traffic Management and Reinforcement Learning in Congested Intersections
http://www.purdue.edu/discoverypark/nextrans/research/research_in_progress.php
Record Type: UTC
The loss of time and resources due to congestion, especially in urban areas, is significant. Appropriately operated traffic signals help to smooth the flow of traffic, leading to a reduction in commute time and fuel consumption. This study seeks to develop an agent-based traffic management technique with reinforcement learning principles. Agents, working independently within the same network, will learn from their environments to minimize travel time and reduce stoppage. The information produced by this innovative research will be applicable to improvements in mobility and reliability in the region.
Start date: 2010/10/1
Status: Active
Total Dollars: $117,786.00
Source Organization: Purdue University, West Lafayette
Date Added: 06/16/2011
Index Terms: Traffic congestion, Highway traffic control, Traffic flow, Intersections, Travel time, Fuel consumption, Agent based models, Traffic signal timing, Mobility,
|
|
NEXTRANS
http://www.purdue.edu/discoverypark/nextrans/
Purdue University, West Lafayette
3000 Kent Avenue
Lafayette, IN 47906
USA
Phone: (765) 496-9729
Fax: (765) 807-3123
|
|
Benekohal, Rahim F.
Phone: (217) 244-6288
Email: rbenekoh@uiuc.edu
|
