The combination of continuous network design and route guidance

Highlights

• A traffic management measure combining the route guidance and the continuous network design is presented.

• A path-based algorithm is given to solve system optimal problem with user constraints.

• The simulated annealing algorithm is improved to solve the bi-level problem.

• Results show that system performance can be improved with the route guidance measure of this paper.

Abstract

In this study, a traffic management measure is presented by combining the route guidance of Advanced Traveler Information System (ATIS) and the continuous network design (CNDP) to alleviate increasing traffic congestion. The route guidance recommends the travelers to choose the shortest path based on marginal travel cost and user constraints. The problem is formulated into a bi-level programming problem. The most distinct property of this problem formulation is that the feasible path set of its lower-level problem is determined by the decision variable of upper-level problem, while in conventional transportation network design problems the feasible path set for lower-level traffic assignment problem is fixed to be all the viable paths between each specific origin-destination pair. The simulated annealing algorithm is improved to solve this bi-level problem. A path-based traffic algorithm is developed to calculate the lower-level traffic assignment problem under the route guidance. Compared to the results of conventional CNDP, the measure presented in this study can better improve the transportation network performance.

Introduction

In presence of rapid economic development, urbanization and population growth, almost all large cities worldwide in the world are facing the serious problem of traffic congestion. Traffic congestion has induced not only huge economic loss, but also environmental deterioration. Based on the report of Texas Transportation Institute, the congestion bill in the United States alone was $67.5 billion in the year 2000, comprised of 3.6 billion hours of delay and 5.7 billion gallons of gas [25]. International Energy Agency said that 23% of global energy related carbon emission in 2004 are related to transportation [13]. It is reasonable to believe that the world will soon have to confront high levels of air pollution and congestion problems caused principally by the unrestricted use of private cars, and have to deploy practical instruments to achieve transportation sustainability efficiently, effectively and in a politically feasible manner [35].

There are generally two ways to alleviate traffic congestion: increasing traffic supply (capacity) and reducing traffic demand [35]. The former way is usually called network design problem (NDP) in transportation network, which determines the enhancement of existing link capacity or the addition of network candidate links. Generally, NDP can be classified into three classes: CNDP (determining the optimal capacity enhancement for a subset of the existing links and its deterministic variables are continuous), discrete network design problem (DNDP) [30], [22] (dealing with the optimal location of new links addition from a set of candidate links and its deterministic variables often are expressed by 0–1 integer), and mixed network design problem (MNDP) (mixture of the CNDP and DNDP) [18].

However, disparate evidence indicates that the enhancement of road capacity induces a greater volume of traffic [9], [10]. Besides, the limitation of land resources in cities cannot support the unlimited increase of link capacity to solve traffic congestion. Basically, more sustainability issues should be considered in NDP [28]. The other measure to reduce traffic congestion is demand-oriented strategies or demand management. Historically, congestion pricing as a demand management instrument has been paid much attention both theoretically and practically. However congestion pricing is perceived as a flat tax since it requires the travelers to pay more for using public urban infrastructure. Meanwhile, there are equity debates of the congestion pricing. So congestion pricing causes the general political resistance and is only applied on urban road in a few cities worldwide [35]. Other than congestion pricing, some quantity control methods to reduce traffic demand are also applied in practice. For example, rationing policies on vehicle usage are used in Mexico City [5], Beijing and Guangzhou, China [11]. Under short-term ration of vehicle usage, observable congestion reduction and air quality improvement have been reported. But it may lose its effectiveness over time as car ownership increases (e.g., there is evidence that driving restrictions in Mexico City led to an increase in the total number of vehicles [5]).

Some researchers also study the combination of NDP and traffic demand management. For example, Wang et al. [32] considered the combination of CNDP and a tradable credit scheme and proved its effectiveness to improve traffic congestion by numerical examples. In this paper, we will study the combination of CNDP and route guidance.

With the development of Intelligent Transport System and advanced techniques of information in the past decades, the advanced traveler information system (ATIS) can easily provide travel information or give travel recommendations. It is widely believed that route guidance information to the travelers is able to efficiently reduce traffic congestion and enhance the performance of traffic networks [34]. Nowadays a large portion of the private cars have been equipped with ATIS devices. While the prices of those deices keep going down, many more travelers are likely to use them and rely on route guidance to achieve trips in the near future. Therefore, it is imperative for the transportation authority to understand how to incorporate the route guidance into the transportation network design so that the network performance is optimized. Traditional network design problems in the literature have not considered the route guidance, assuming that travelers follow user equilibrium (UE) principle to minimize their individual travel costs. In this study, we assume that when transport planners decide to improve the road network, they have to consider that route guidance information would be provided to the travelers and therefore the resultant network traffic flow pattern is different from the UE traffic assignment. Besides, noting that the simple and naive system-optimal based route guidance is subject to unfairness issue, we assume that route guidance strategy with certain user constraints is applied to reduce the unfairness. Indeed, the network traffic flow pattern achieved with this route guidance strategy is constrained system optimal (CSO). The problem studied in this paper, i.e., the combined continuous network design and route guidance, is then formulated into a bi-level programming. A modified simulated annealing algorithms are proposed to solve the problem. To summarize, the main contribution of this research work is to fill in the research gap in transportation network design problems by considering the route guidances of the traveler information system.

The paper is organized as follows: Section 2 presents a bi-level programming problem to model the combination of CNDP and route guidance. The algorithm to solve the bi-level programming problem is given in Section 3. Section 4 gives the numerical test and the conclusions of the study is presented in Section 5.