In this paper, we propose a dynamic route-exchanging mechanism based on anticipatory stigmergy and demonstrate its efficiency. Next-generation traffic management systems based on probe vehicle data have been attracting attention. Ito et al.[2] [3] [4] previously proposed a traffic management method based on anticipatory stigmergy that can search for an alternative route to avoid expected congestion by sharing the probe vehicles’ expected locations in the near future, and they found that anticipatory stigmergy works well in particular experimental settings. On the other hand, Takahashi et al[6] identified two key issues: (1) The oscillation of congestion occurs because drivers do not know other drivers’ decision making. This problem is well-known as the El Farol Bar Problem or the congestion game. (2) The saturation level of navigation systems could affect the performance of a dynamic route exchanging-mechanism. In this paper, we propose a new dynamic route-exchanging mechanism that can address the above two issues. In the basic procedure, each vehicle submits its intention about its near-future position (60 seconds). Then the traffic management center computes the near-future congestion information for each link. This information is anticipatory stigmergy. If there is an over-congested link after 60 seconds, the vehicles assumed to come to those links are allowed to negotiate with each other so that some of them will change their near-future routes. In this mechanism, vehicles automatically negotiate based on their rational judgment on the trade-off between travel time needed for passing the assigned route and the ”concession coefficient” that represents how a driver can concede the way. The experimental results demonstrate that our new route-exchanging mechanism performs well for the efficiency of traffic flow when the saturation level of probe vehicles is greater than 70%.
[1]
Yosef Sheffi,et al.
Urban Transportation Networks: Equilibrium Analysis With Mathematical Programming Methods
,
1985
.
[2]
Ivan Marsá-Maestre,et al.
The Comparison of Stigmergy Strategies for Decentralized Traffic Congestion Control: Preliminary Results
,
2012,
PRICAI.
[3]
Takayuki Ito,et al.
Evaluation of anticipatory stigmergy strategies for traffic management
,
2012,
2012 IEEE Vehicular Networking Conference (VNC).
[4]
Takayuki Ito,et al.
Evaluation of Traffic Management Strategies with Anticipatory Stigmergy
,
2014,
J. Inf. Process..
[5]
Mitsuru Ishizuka,et al.
PRICAI 2012: Trends in Artificial Intelligence
,
2012,
Lecture Notes in Computer Science.
[6]
Takayuki Ito,et al.
Stability Evaluation of Route Assignment Strategy by a Foresight-Route under a Decentralized Processing Environment
,
2013,
2013 IEEE/WIC/ACM International Joint Conferences on Web Intelligence (WI) and Intelligent Agent Technologies (IAT).
[7]
Luca Maria Gambardella,et al.
Ant colony system: a cooperative learning approach to the traveling salesman problem
,
1997,
IEEE Trans. Evol. Comput..