Abstract:
Traffic congestion is one of the most important problems we face in our daily life. This
problem has a huge effect on economy and environment. The total of congestion cost in
America in 2010 reached to 101 billion dollar and it is expected to reach to 175 billion
dollar in 2020 [1]. Despite of many research in this field the problem still exists all over
the world.
The new domain for solving this problem is using agent based technology, which can
model the traffic network as agents, these agents can communicate, cooperate to achieve
the optimal performance of the network.
Our thesis proposed a solution for this problem based on agent technology. We use
various agents in our model which include lane agent, vehicle agent and traffic light agent.
We use lane agent for every lane of the network to compute the weight of the lane and
send it to traffic light agent. This weight depends on a new model called horn model. The
horn is the sound produced by vehicles to give an indication that green sign is needed,
this sound increase with waiting time and decrease as the distance of waiting vehicles
from traffic light increase .
. Traffic light agents use the congestion factor and weight it receive from lane agents to
optimize traffic light signal. Optimization is done by giving the needed time for every lane
according to the number of waiting vehicles and to keep track of the waiting time of other
vehicles in other lanes by using lane weight. This solution can give a dynamic cycle time
according to real time traffic data. Route guidance system based on congestion factor is
used also to stop sending vehicles to congestion area. Traffic light agents communicate to
ii
Description:
no of pages 70, 27263, informatics 5/2013 , in the store
