China
Africa
Explore chapters and articles related to this topic
Dynamic Road Management in the Era of CAV
Published in Hussein T. Mouftah, Melike Erol-Kantarci, Sameh Sorour, Connected and Autonomous Vehicles in Smart Cities, 2020
Mohamed Younis, Sookyoung Lee, Wassila Lalouani, Dayuan Tan, Sanket Gupte
Another group of AIM considers multiple intersections in a cooperative way [112,113,114]. In [113], a study is reported on how a single AV may cross multiple signalized intersections without stopping in a freeflow mode; an optimal eco-driving algorithm is proposed to generate the acceleration and speed profile by considering multiple intersections jointly rather than dealing with them individually. The multiintersection control is modeled using multiple agents across a network of interconnected intersections. From the multiagent perspective, AVs dynamically modify their scheduled paths based on different navigation policies and in response to minute-by-minute traffic conditions. Therefore, for a large road network, an instance of Braess’ paradox may be experienced where opening additional travel options for the vehicles reduces the efficiency of all vehicles in the system. Such paradox is handled in [112]. Meanwhile, Li et al. [114] have developed an intersection automation policy (IAP) for serving requests for green light made by both AV and human-driven vehicles. IAP exploits real-time tracking of vehicle location to predict arrival at intersections along its route where requests for green signals are anticipated. A schedule for green time is then devised based on the phase-time-traffic hypernetwork model, articulated in Figure 5.4, which represents heterogeneous traffic propagation under traffic signal operations. Thus, the signal time and vehicle movements are optimized for all vehicle types.
Traffic Network Modeling
Published in Dušan Teodorović, The Routledge Handbook of Transportation, 2015
Srinivas Peeta, Henry Liu, Xiaozheng He
This section presents the well-known Braess paradox (Braess, 1969; Braess et al., 2005) to illustrate the difference between the concepts of user equilibrium and system optimum. By intuition, we expect that building more roads can enhance the highway system performance. However, the Braess paradox shows a contradiction in which the addition of a link causes the increase of average travel cost.
Introduction
Published in Joseph Y.-T. Leung, Handbook of SCHEDULING, 2004
Since the prior discussion shows that the coordination ratio may be very large or even unbounded, it is interesting to study the question of how to design networks to ensure that the coordination ratio is small, that is, for a given class of latency functions, how to modify a given network to obtain the best possible flow at Nash equilibrium. It is well-known that removing the edges from a network may improve its performance. This phenomenon was first shown by Braess [41] (see also [42]) and is commonly known as Braess’s Paradox, see Figure 42.2. Roughgarden [43] gave optimal inapproximability results and approximation algorithms for several basic network design problems of the following type: given a network with n nodes, edge latency functions, a single-source pair, and a rate of traffic, construct a subnetwork minimizing the flow cost at Nash equilibrium. For networks with continuous, nonnegative, nondecreasing latency functions, Roughgarden showed that on one hand, there is no (n/2-ϵ)-approximation polynomialtime algorithm unless P=NP, and on the other hand, there is an n/2-approximation polynomialtime algorithm for this problem; for linear latency cost functions, the bound of n/2 should be replaced by 43. Moreover, Roughgarden [43] showed that essentially the following trivial algorithms achieve the optimal bounds: given a network of candidate edges, build the entire network. For more discussion about this problem we refer the reader to [43]; research on a related model has been recently reported in [10].
Connectivity of two-dimensional assemblies: trusses and roads
Published in Civil Engineering and Environmental Systems, 2021
It should be noted that there are other criteria for identifying the weakest link. For example, the metric ‘spanning edge betweenness’ corresponds to the fraction of shortest paths between all nodes that contain the edge in question and reflects the importance of that edge to maintaining the network (Teixeira, Santos, and Francisco 2016). The concept of ‘link entropy’ has been used to quantify the significance of the edge for maintaining global connectivity and is based on the information entropy of the two nodes it connects; where the information entropy of a node is a measure of the degree to which it belongs to different clusters (Qian et al. 2017). Other measures such as the Alpha index, Beta index, and Shimbel distance can also be found in the literature (Ducruet and Rodrigue 2020). Measures can also be based on the effect of adding, rather than removing an edge (Chen et al. 2019). In this context, Braess’s paradox highlights the fact that the addition of a link can actually decrease network performance. While originally demonstrated in the context of road networks (Steinberg and Zangwill 1983), the paradox can apply to electric power grids too (Blumsack et al. 2007). We have chosen Newman’s (2004) edge betweenness criterion however, because it is one of the simplest, and – given that our exploration of connectivity in roads networks and structural trusses is rather tentative – probably the most fit for purpose.
Effect of reducing the price of anarchy on fairness in highway resource allocation for individual users
Published in Transportmetrica B: Transport Dynamics, 2019
The result that the value of is bounded when the total flow is large implies that less route switching will take place under high demand when a system moves from the UE state to the SO state. In other words, when the total demand becomes very large, the SO flows are similar to the UE flows and thus yield the same result. The effect becomes more pronounced when is large. This property of the UE and SO relationship can be utilized to explain the surprising results obtained by Pas and Principio (1997) for the Braess’s Paradox problem (Braess 1968). In their work, they considered a broad range of demand and examined the effect of the demand level on the occurrence of Braess’s Paradox on a simple network. They revealed that the network would grow out of Braess’s Paradox when the total demand is sufficiently high. The outcome of their research might be explained by the merge of the UE and SO states for high demand shown in the above analysis since the amount of the flow transfer between different routes for achieving the SO state is a decreasing function of total demand. Whether this result is realistic or not is yet to be further examined since the result is indeed quite counter intuitive. After all, the underlying models used in this paper and in Pas and Principio (1997) are static models. Models considering time-dependent queues would yield different results (Lin and Lo 2009).
A microsimulation based analysis of the price of anarchy in traffic routing: The enhanced Braess network case
Published in Journal of Intelligent Transportation Systems, 2021
Aleksandr Belov, Konstantinos Mattas, Michail Makridis, Monica Menendez, Biagio Ciuffo
Because of the highly detailed modeling approach and difficulties to calculate SO using a microsimulation model we start with a simple network topology—a well-known 5-link network from the Braess paradox example. Braess paradox is the counterintuitive situation when the addition of a new link to the network decreases its efficiency in the UE state (Braess, 1968). One could argue that it is not a true paradox because there is a theoretical explanation for it. In any case, this is a typical example of PoA appearance since the situation can be improved by rerouting traffic to avoid the usage of that link. The network and the simulation tests were designed and implemented using the Aimsun commercial simulation software (Aimsun Next, n.d.).