Multi-agent based Simulation of traﬃc in
The Duy Bui, Duc Hai Ngo, Cong Tran
College of Technology
Vietnam National University, Hanoi
Abstract. There is always need for a good simulation for traﬃc in Vietnam in order to help transportation planners to improve the current traﬃc system. Over recent days, there has been severe traﬃc congestion in many streets of big cities in Vietnam such as Hanoi and HoChiMinh city. There is an urgent need for measures to deal with increasing congestions. The simulation of traﬃc in Vietnam is a hard problem due to two main reasons including: (1) the traﬃc participants in Vietnam do not give way according to the rule; (2) the participants do not consider that when waiting for the vehicles in front, stopping in the intersection is obstructing the traﬃc ﬂow. In this paper, we propose a multi-agent based simulation system for traﬃc in Vietnam to help transportation planners to ﬁnd treatments to the problem of congestion of the traﬃc system in Vietnam as well as to test new designs before committing resources to actually building the transportation infrastructure. By allowing a user to design diﬀerent road systems as well as to create diﬀerent simulation scenarios with diﬀerent agent proﬁles, our system can simulate the dynamic of traﬃc in Vietnam in diﬀerent situations.
Keywords: Traffic simulation, Multi-agent based simulation
For all countries, transportation is one major part of infrastructure, which has a direct impact on the economic and social development of the country. However, building a good transportation system is always a diﬃcult problem for every country in the world. In developed countries, the development of transportation system always requires huge expenses. In addition, transportation planners always have to have a strategic vision, which can identify a clear plan to develop the transport system. In Vietnam, transport infrastructure does not keep up with the social and economic development, thus obstructing the modernisation and industrialisation of the country. Currently in Vietnam, traﬃc is a sore issue for policy makers. In cities such as Hanoi, the transport system is chaotic, due to narrow road, increasing number of vehicles, and lack of consciousness to follow the traﬃc rules from participants. The result is an increase in congestion, accidents and environmental pollution. So the search for solutions to develop transport systems in Vietnam is still a challenging question.
In many countries, the traﬃc simulation system on a computer has long been studied for ﬁnding and evaluating solutions to develop transport systems [1, 2, 4, 7, 8, 9, 10, 11]. Before blocking a road, opening a new route or timing the green light and red light in the real world, the developers can test the system in the simulation system. Either for the purpose of modeling short-term traﬃc dynamic on a single road section or for transportation planning describing behavioral pattern in a network on a larger time scale, existing simulation systems often assume that the behaviour of drivers involves accelerating, braking, and changing lanes. This results in a range of methods from mathematical models, macroscopic models to nanoscopic models .
The traﬃc in Vietnam, however, is diﬃcult to model due to the more complicated behaviour of drivers. As the drivers do not follow the lanes, the traﬃc simulation becomes closer to the crowd simulation. Therefore, mathematical models or macroscopic models are diﬃcult to obtain in this case. Instead, microscopic models are more suitable. In this paper, we propose a microscopic approach of using multi-agent based simulation for the traﬃc in Vietnam. The driving behaviour of an agent in our system is characterized by diﬀerent parameters such as planning time, maximum speed, and acceleration rate. By allowing a user to design diﬀerent road systems as well as to create diﬀerent simulation scenarios with...
Please join StudyMode to read the full document