[分享]关于科学杂志中的交通流理论文章

lizhi

[这个贴子最后由水寿松在 2004/09/10 06:06pm 第 1 次编辑]

关于科学杂志中的交通流理论文章

1.1TRAFFIC THEORY:
《Jams, Waves, and Clusters》  Dirk Helbing and Martin Treiber*
Have you been suffering from traffic jams lately and asking yourself why freeways are not free ways anymore? Help may be on the way. There have been several recent advances in traffic theory, notably those that treat traffic like a fluid. Researchers at Seoul National University (1) have now offered an interpretation of a recently discovered state of congested traffic, called "synchronized" traffic (2). Their fluid-dynamic simulations could be a useful tool for an optimization of traffic flow on motorways.
When Nagel and Schreckenberg presented their cellular automaton model of traffic flow in 1992 (3), allowing for a more than real-time simulation of the entire road system of large cities, they probably did not anticipate the resulting flood of publications and the enthusiasm among scientists on the subject of traffic theory. By treating huge numbers of interacting vehicles similar to classical many-particle systems, physicists have recently contributed to a better understanding of traffic flow. The mathematical tools that they use, stemming mainly from statistical physics and nonlinear dynamics, have proved their interdisciplinary value many times. This includes concepts reaching from self-organized criticality and phase transitions up to the kinetic theory of gases, fluids, and granular media.
In traffic, drivers try to maximize their own advantage (that is, their velocity, safety, and comfort) within the constraints imposed by physical limitations and traffic rules. Under certain conditions, their competitive, nonlinear interactions give rise to the formation of collective patterns of motion like traffic jams. The various observed phenomena on freeways are surprisingly rich: Apart from free traffic and extended traffic jams behind bottlenecks, there are localized clusters (small moving jams) and stop-and-go waves. In addition, Kerner and Rehborn (2) have recently discovered a hysteretic phase transition from free traffic to a new form of congested traffic (mostly appearing close to on-ramps) that had not been identified in more than 40 years of traffic research, they say. Kerner and Rehborn call it "synchronized" traffic because of the synchronization of velocities among lanes. However, the more characteristic feature is its high flow in spite of the breakdown of velocity, which is in contrast to typical traffic jams. Downstream of the ramp, the breakdown of velocity eventually relaxes to free traffic in the course of the freeway. Another interesting property is the wide scattering of synchronized traffic states, when plotted in the flow-density plane, which differs from the quasi-linear density dependence of free traffic flow.

Life in the slow lane. Formation of the recurring hump state (RH) on a freeway for the model and parameters used by Lee et al. (1). Because we used free rather than periodic boundary conditions, the inflow at the on-ramp located at 0 km did not need to be balanced by an off-ramp, which is more realistic. The constant ramp flow of 318 vehicles per hour and lane causes a higher vehicle density downstream of the ramp. Nevertheless, free flow pertains, until a short perturbation of the ramp flow occurs at time 0 min, which moves downstream in the beginning. However, with growing amplitude, the perturbation changes its propagation speed, reverses its direction, and finally induces another, bigger perturbation, when passing the ramp. This process repeats again and again, in this way generating the oscillating RH state. When passing the ramp, the perturbations continue their way upstream, until they merge with one of the humps that were born later.

Lee et al. (1) have now suggested an explanation for this hysteretic phase transition. They simulated freeways, including on- and off-ramps, with a fluid-dynamic traffic model that is closely related to the Navier-Stokes equations for viscous, compressible fluids. However, it contains an additional relaxation term describing adaptation of average vehicle velocity to an equilibrium velocity, which monotonically decreases with growing density. In comparison with previous simulation studies, Lee et al. used another velocity-density relation and a considerably different set of parameters. With a temporary peak in the on-ramp flow, they managed to trigger a form of oscillating congested traffic that is propagating upstream, but pinned at the location of the ramp (see figure above). They call it the "recurring hump" state (RH) and compare it with autocatalytic oscillators in chemistry. Free traffic would correspond to a point attractor and the oscillating traffic state to a stable limit cycle. In terms of nonlinear dynamics, the transition corresponds to a Hopf bifurcation, but a subcritical one, because the critical ramp flow depends on the size of the perturbation.
Lee et al. point out that free traffic survives the assumed pulse-type perturbation of finite amplitude, if the ramp flow is below a certain critical value. However, once a RH state has formed, it is self-maintained until the ramp flow falls below another critical value that is smaller than the one for the transition from free traffic to the RH state. This proves the hysteretic nature of the transition. Moreover, Lee et al. could show the gradual spatial transition from the RH state to free flow downstream of the ramp. They also managed to reproduce the synchronization among neighboring freeway lanes as a result of lane changes. Therefore, they suggest that their model can describe the empirically observed first-order phase transition to synchronized traffic. The two-dimensional scattering of synchronized traffic states is understood as a result of the fact that the amplitude of the oscillating traffic state depends on the ramp flow.

On the highway. Phase diagram of the various traffic states that can occur close to an on-ramp in the presence of small perturbations in the ramp flow. We show the dependence of the traffic states on the ramp flow and the ramp length for a flow of 1800 vehicles per hour and lane on the freeway. For small ramp flows, free traffic (FT) survives. At higher inflows, two different kinds of RH states can build up, either triggered stop-and-go waves (TSG) or oscillatory synchronized traffic (OST). High ramp flows are associated with a homogeneous form of synchronized congested traffic (HST).

Although the interpretation of synchronized traffic by Lee et al. does not quantitatively agree with the observations, in various respects it comes pretty close to reality. Meanwhile, our recent work has offered a more complete explanation (4). Above all, the findings are also of great practical importance. A detailed analysis shows that there is a whole spectrum of different states that can form at ramps. Their occurrence decisively depends on the inflow as well as the ramp length (see figure above). This is relevant not only for an appropriate dimensioning of ramps, but also for an optimal on-ramp control.
Traffic theory is now more interesting than ever before. Recent advances have yielded a better understanding of traffic flow phenomena as well as realistic and fast simulation models. Scientists are now prepared to design on-line controls for efficient traffic optimization, calculate the environmental impact of congestion, and develop methods for traffic forecasts.
References
1.H. Y. Lee, H.-W. Lee, D. Kim, Phys. Rev. Lett. 81, 1130 (1998) [APS].
2.B. S. Kerner and H. Rehborn, ibid. 79, 4030 (1997) [APS].
3.K. Nagel and M. Schreckenberg J. Phys. I France 2, 2221 (1992).
4.D. Helbing and M. Treiber, Phys. Rev. Lett. 81, 3042 (1998) [APS]; preprint available at http://xxx.lanl.gov/abs/cond-mat/9809324..

good

[这个贴子最后由水寿松在 2005/03/19 05:26pm 第 1 次编辑]

首先建议把该贴的题目改成《Jams, Waves, and Clusters》。
另外，文章中多次写着see figure above，lizhi坛有可否传图片，以加深对文章的理解。
我翻译了文章的最后两段，不知道理解的准确否，请大家指正。
Although the interpretation ……
尽管Lee et al.阐述的同步交通流在数量上与观测不符，在很多方面它更好的接近了事实真相。其间，我们当前的工作提供了一个更完整的解释。当然，发现也是最重要的。详细地分析显示在不同状况下坡道形成的同步交通流有一个完整的图谱。他们的发生依赖于入流，也就是说坡道的长度。这不但关系到坡道适当的尺寸而且关系到坡道的最佳控制。
Traffic theory ……
交通流理论比以前更令人关注。最近的发展已经可以更好的理解交通流现象，也就是建立了真实可靠的仿真模型。科学家们正准备设计在线交通最优化控制，计算交通阻塞的环境影响，并且发展交通预测模型。

这样的工作需要鼓励和支持，＋100金币

itslu

我觉得不一定象楼上说的把帖子的题目改成《Jams, Waves, and Clusters》，大家都知道Nature在学术界的地位，我想楼主起的题目也是想让大家知道在Nature这样高水平的杂志上也有交通流方面的文章，这从一个侧面体现了交通工程学科研究是一门科学，是需要去研究的。
不知道从Nature创刊以来，交通方面的文章是不是就这一篇？我也一直在查找进NAtrue数据库的用户名和密码，不过没有找到。不知道哪位能找到影印版的Nature杂志和Science杂志，扫描一下交通方面的文展上传上来啊！多谢！

hxwfirst

把交通学科事业发扬光大!!