论文总字数：56338字

摘 要

收费道路是一种新兴的城市基础设施，并且越来越多的出现在澳大利亚各个城市。然而，出行者回避收费道路的行为却可能带来不良的影响，例如产生噪音，排放加剧，事故增加，这些不良影响都增加了社会成本及环境成本。城市物流是一种基于系统方法论对货物进行分拨运输的综合方法。它提倡利用创新的方案规划城市内的物流，以减少系统总成本（包括经济、社会和环境成本）。

然而，目前在墨尔本，过路费定价主要是由车辆行驶距离和车型来决定的，现有定价下的过路费能够带给出行者和社会多少益处尚不清楚。本文建立了成本分析模型，估算了在不同过路费定价水平下，承运人出行产生的成本，以观测当该收费水平变化时，对车辆路径选择行为以及拥挤程度有何影响。同样，收费定价对社会及环境成本的影响也考虑在内。

最后，通过案例分析,利用模型证明了如今墨尔本现有的道路收费定价虽然能够鼓励出行者选择收费道路，有效减少车辆产生的负外部性，但是不足以偿还车辆使用收费路时产生的负外部性。过路费上调后产生的影响也考虑在内。本文仅考虑了基于不同车型的单个车辆的成本。因此后续工作需要收集车辆总数和不同车型所占百分比等数据以计算每次出行产生的总成本。

关键词：收费道路；出行成本；城市物流; 路径选择。

Content

1. Introduction 1

2. Development of Database 3

2.1．Truck cost concept 3

2.2．Vehicle classification 4

2.3．Vehicle Operating Cost (VOC) 5

2.4．Value of Time (VOT) 6

2.5．Environmental Cost (EC) 7

2.6．Social Cost (SC) 8

3. Analysis and Discussion 8

4. Conclusion and Future Works 13

Acknowledgements 14

References 14

Appendix A 16

1. Vehicle Operating Cost 16

1.1. Distance-related Cost (C_d) 16

1.2. Time-related Cost (C_t) 19

2. Environmental Cost (EC) 21

3. Social Cost (SC) 22

Appendix B 22

Figure 1.Technology route 3

Figure 2.The structure of Vehicle Operating Cost 5

Figure 3.Speed versus Fuel Cost Curve 6

Figure 4.Alternative routes selected for analysis 9

Figure 5.Relationship between decision making probabilities and operator benefit 10

Figure 6.Externalities produced on Freeway and Highway by vehicle type at selected situation 11

Figure 7.Total externalities under different toll levels 12

Table 1.Vehicle classification 4

Table 2.Vehicle Operating Cost at different speed per Kilometer by vehicle type ($) 6

Table 3.Hourly wage for drivers by vehicle type ($) 7

Table 4.Emission Cost by vehicle age ($) 7

Table 5.Environmental Cost on Highway and Freeway per Kilometer by vehicle type ($) 8

Table 6.Social Cost on Highway and Freeway per Kilometer by vehicle type ($/km) 8

Table 7.Travel Times on Freeway and Highway in different time of a day 9

Table 8.Sum of VOCamp;VOT and Externalities on Freeway and Highway compared with toll level under selected situation ($) 11

Table 9.Comparison between 3 different toll levels and difference in direct cost 12

Table 10.The proposed new toll charge by vehicle type 13

Table 11.Coefficients of Fuel Cost by vehicle type 17

Table 12.Coefficients of Tyre Cost by vehicle type 17

Table 13.Coefficients of Maintenance Cost by vehicle type 18

Table 14.Coefficients of Distance-related Depreciation Cost by vehicle type 19

Table 15.Coefficients of Interest Cost by vehicle type 20

Table 16.Data of Administration Cost by vehicle type 20

Table 17.Data of Registration Cost by vehicle type 20

Table 18. Data of Insurance Cost by vehicle type 21

Table 19.Coefficients of Time-related Depreciation Cost by vehicle type 21

Table 20.Coefficients of Environmental Cost by vehicle type 22

Table 21.Coefficients of Social Cost by vehicle type 22

Introduction

City Logistics involves an integrated approach for managing urban freight (Thompson and Taniguchi, 2001). When evaluating urban freight policies consideration a broad range of impacts should be considered including environmental and social as well as economic (Thompson, 2015). It is also important to consider the effects for a number of stakeholders, including residents, carriers and receivers. A variety of supply, demand and impact models are necessary for predicting the effects of urban freight management measures (Taniguchi and Thompson, 2002; Taniguchi and Thompson, 2003).

City logistics has been defined as, “the process for totally optimizing the logistics and transport activities by private companies with support of advanced information systems in urban areas considering the traffic environment, the traffic congestion, the traffic safety and the energy savings within the framework of a market economy” (Taniguchi et al., 2001). From this definition we could know that how to balance the economic benefits with environmental and social optimization is the key point in city logistics. Maroudas-Tsakyrellis (2011) paid attention to how to achieve Green City Logistics, aiming at reducing the negative effects from emissions. They proposed that two future development trends of city logistics in Stockholm: Sustainability and Viability. Taniguchi E (2014) proposed three measures towards making cities more sustainable based on the improvement of city logistics: applying ICT and ITS system, training logistics managers and using public-private partnerships to promote city logistics policy measures.

It was found that total externalities approximately produced by freight vehicles in Australia are about 28.2 billion dollars which is 1.9% of the GDP in 2014 (1,454 billion dollars). This consists of $12.8 billion of environmental costs and $15.4 billion of social costs. The total kilometres travelled by freight vehicles in Australia were obtained from the Australian Bureau of Statistics database, 2014. However, it was assumed that 40% of vehicles would choose freeways and rest, 60%, choose highway at an average speeds of 95km/h and 45km/h, respectively. This fundamental study shows that how significant the externalities are and which needs to be beared by society as a whole.

This paper mainly focuses on trucks for considering city logistics. In a report by BTRE, the Australian domestic freight task measured 521 billion tonne kilometres in 2007, 35% of that was conveyed by road. Compared with that, railway is taken on 40% of the freight while ship carried 25%. Less than 0.1 per cent is carried by air (Bureau of Transport and Regional Economics, 2006). BTRE also predicted that road freight volumes will become double by 2030. Road freight plays a major role in the Australian economy as well as influencing city logistics. In Mexico, a research of how city logistics measures can improve environment and society have been conducted. (Jaller M et al., 2016) They considered several methods including reducing travel distance and travel time and find that those methods can lower carbon emissions, alleviate traffic congestion, reduce the frequency of accidents and encourage off-time delivery. It affirms that encouraging road users to use toll road can be a powerful tool to reduce the transport externalities due to that toll road usually owns shorter distance and less travel time than highway and arterial. Thus, we need to get a deeper understanding of road-user charge.

Road-user charges have long been heatedly debated in the field of transport economic. Congestion pricing can efficiently reduce the large social and environmental cost imposed by the heavy traffic such as time consumption, accidents, high noise level, and emissions, but there are also resistances in its implementation. C. Robin Lindsey and Erik T. Verhoef (2000) reviewed the economic theory of congestion pricing from an economic point of view and affirm the roles of toll in traffic demand management. In contrast, Evans A W (1992) proposed two defects of road pricing in basis of practical application. In his point of view, the first disadvantage is that road pricing may benefits residents more than its object users that can cause social inequity. The other disadvantage is that road pricing lead to a significant revenue generation for government, in which the motivation of toll policy promotion may be unjustifiable (Evans A W, 1992).

Road toll is often regarded as an effective instrument to alleviate the traffic congestion. Toll roads work by diverting traffic flows and thus they are able to send peak-hour traffic to off hour or other links. (Fan W, 2016)But improper toll level also causes side-effect (Gronau R, 1999).For example, avoidance of toll roads by trucks can cause increased social and environmental costs. Higher levels of noise, emissions, fuel consumption and crashes can result from trucks using alternative routes. Therefore, the optimal toll level need to be considered.

The following efforts have made to research this problem from a theoretical perspective. From a economic point of view, C. Robin Lindsey and Erik T. Verhoef (2000) reviewed the economic theory of congestion pricing, they also conclude the obstruction to implement the toll policy. Reuben Gronau (1999) made a discussion about a single toll road in a non-toll environment and emphasized the functions of the toll. A bi-level genetic algorithm (GA) based approach were proposed to finding the optimal toll locations and toll levels simultaneously in a multiclass network (Wei Fan, 2016). In that paper, the author state the concept of first best pricing which is that marginal-cost toll is charged on each link to achieve a user equilibrium flow pattern to achieve a system optimum at total network level .But due to the high cost of setting devices on every link to collect toll and low public acceptability of toll, it is impossible to achieve first best pricing. Therefore, we should pay more attention on second-best pricing scheme. In this scheme, only a subset of links is chose for charging (not all links). There is a paper deals with the second-best pricing problem, using a binary genetic algorithm to search optimal toll locations dynamically and a simulated annealing method to search optimal toll levels. (Yang H, Zhang X, 2003)

Improper toll charges may have mislead the freight route selection. Chen M, Bernstein D H(2004) considered different characteristics on road users(eg. income levels) ,and developed a model to evaluate toll impact to what extent on various user groups route choice behaviours ,aiming at finding the optimal toll on link to get the lowest total cost . Besides, there are not enough studies about toll impacts in term of saving total cost. In this respect, which is the importance of the research.

The main objective of this paper is to find out the effective toll charge for freeways for freight vehicles to minimise total cost which includes environmental and social cost. It is worth mentioning that truck cost is the only factor need to be considered when carriers deciding driving route (Reuben Gronau, 1999).So we have to find out various cost components for freight vehicles(direct cost and indirect cost) at first, trying to establish a database about different cost. Finally we could trade-off between using alternative routes for freight transport, by making a comparison of truck cost analysis between toll road and toll-free road is presented to see how toll levels effects users’ route choice behaviour.

The rest of the paper is organized as follows. Section 2 introduces the method to determine toll charge which is the development of database and models of truck cost analysis. Section 3 shows a sample using truck cost models in Melbourne between a toll road (freeway) and an arterial road in a same O-D pair, and gives discussion. Section 4 makes conclusion and gives suggestions for toll charge, finally proposes future work. Following figure shows the technology route of this paper.

Figure 1.Technology route

2. Development of Database

Before we present the truck cost analysis, the content of Road User Cost (RUC) and Vehicle Operating Cost (VOC) need to be distinguished.

2.1．Truck cost concept

Analysing the Road User Cost (RUC) is a crucial part when starting a new road project. RUC is usually used to conduct Benefit-Cost Analysis (BCA) and Life-Cycle Cost Analysis (LCCA) during traffic planning. (Qin X, Cutler C E, 2013) For traffic managers, the specific roles of RUC mainly include: (1) Predicting the long-term effects of road construction on traffic public. (2) Designing economic measures to deal with constructors. (3) Deciding the proper order of a project. (4) Employing the limited funds to conduct road maintenance. (Qin X, Cutler C E, 2013; Ihs A, Gustafsson M, Eriksson O, et al,2011).

According to the newest HDM-4 manual (Highway Development and Management Model), RUC comprised Vehicle Operating Cost (VOC), Time Cost (usually convert to a monetary value of time), Crash Cost and Environmental Externality Cost (emission and noise cost) (Bennett, CR amp; Greenwood, 2006).

A software method based on RUC calculation is put forward namely TRIPS (TRansport Improvement Programming System), in which the inputs are congestion levels, trip volumes as well as some road characteristics including road roughness, grade and curvature. (LLOYD B, TSOLAKIS D, 2001) The HDM-4 (Bennett amp; Greenwood 2006 and Stannard amp; Wightman 2006) models are of an empirical kind, with the coefficients derived by the statistical analysis of observations. (Naude C, Toole T, McGeehan E, et al, 2015) Hassan R, Martin T, Thoresen T, et al (2008) introduced a RUC model specific to High Productivity Vehicle by Road pavement type and stated which effects have heavy vehicles caused on road pavement from the benefit and cost perspectives.

The Vehicle Operating Cost (VOC) is a significant component in RUC. VOC is the total cost of road transport which consist of fuel, tyre, parts, labour, oil and capital costs. It may also include travel time and crew costs. (Bennett, CR amp; Greenwood, 2006) ByDrewello H, Scholl B (2015) and Combes P P, Lafourcade M (2005), VOC is a combination of time-related costs and distance-related costs. The cost by time is generated when the vehicle is not running on road, which contains interest, insurance, tax, administration, registration fee and time-related depreciation. In contrast, distance-related cost in an on-road cost, such as cost for fuel, tyre, maintenance and distance-related depreciation.

For VOC calculation models system, they can be classified into four different levels of modelling framework that have been developed (LLOYD B, TSOLAKIS D, 2011),these models are, from lower level to higher level, showed as follows:(1) an ’instantaneous’ model (2) an ‘elemental’ model (3) a ‘running-speed’ model (4) an ’average-speed’ model. Each higher level model is an aggregation from a lower level. The lowest model owns the most accuracy. aaSIDRA is a method of ‘four-mode elemental’ model (Akcelik R, Besley M.2013). It divided a whole drive cycle into four drive behaviour: cruise, acceleration, deceleration and stopped. Fuel consumption varies among those drive behaviours under different traffic conditions (the variables are intersection geometry, traffic control methods and congestion levels). aaMOTION is another instantaneous model in basis of second-by-second trip data. (Akcelik R, Besley M.2013). Tan F, Thoresen T (2012) studied the relationship between VOC and Road Roughness, more specifically, Road Roughness directly affect the cost of fuel, tyre and maintenance.

Fuel consumption is a key point when calculate VOC because it varies with many factors, such as speed, payload, Road Roughness, horizontal and vertical alignment, etc. The book’ Guide to fuel consumption analyses for urban traffic management’ comprehensively summarized and classified fuel consumption models in different areas and in different traffic system scale. (Bowyer D P, Akçelik R, Biggs D C, 1985)

2.2．Vehicle classification

The vehicle classification used in this paper is based on the number of axles, including 9 vehicle classification, which is a combination of the classification in HDM-4 (2002) and Transport SA (2006).This paper solely focus on trucks so vehicles like buses or private cars will not be discussed.

Table 1 shows the vehicle classification and features of each type:

Table 1.Vehicle classification

The structure of Truck cost analysis presented as follows. Four cost components in total cost were considered, namely Vehicle Operating Cost (VOC), Value of Time (VOT), Environmental Cost (EC), and Social Cost (SC).The detailed description of each cost component will be given in following text.

2.3．Vehicle Operating Cost (VOC)

Figure 2.The structure of Vehicle Operating Cost

The approach to evaluate VOC refers to the classification method mentioned above. There are two main cost components in VOC, namely：Time-related cost and distance-related cost. The cost by time is generated irrespective of vehicle running on road or not, which contains interest, insurance, administration, registration fee and time-related depreciation. In contrast, distance-related cost is an on-road cost, such as cost for fuel, tyre wastage, maintenance and distance-related depreciation. Cost by time exclude the driver/labour cost, which has been considered separately in the value of time. Figure 2 depicts the VOC structure considered in this study.

Between the two cost components, the distance-related cost is a function of travel speed for obvious reasons the time-related cost has no clear link with the travel speed. Some of the key data such as market price of vehicles, average working distance per vehicle and vehicle life are from credible department including NGTSM (National Guidelines for Transport System Management) , Austroads, ABS (Australian Bureau of Statistics). Other source are from internet, including interest rate, tyre and maintenance, insurance and registration fee. We tried to pick the newest data as far as possible.

Between the two cost components, the distance-related cost is a function of travel speed for obvious reasons the time-related cost has no clear link with the travel speed. It is worth mentioning that the relationship between speed and fuel cost accords with the following curve. (Figure 3) This curve is a HDM style models based on mechanistic approach under interrupted flow condition. (Austroads 2005).Though it is important to consider road configuration, geometry and layout as well as traffic volume (speed) when calculating fuel consumption, it is difficult to get accurate value of those factors. So model with single variable for average speed was most commonly used. (Naude C et al., 2015) The lowest cost point is 60km/h of the speed. With the growth of speed fuel cost significantly drops down until the speed reaches 60 km/h. After that fuel cost rises slightly. In general, free-flow speed (gt;60 km/h) obviously consumes less fuel than stop-start speed (lt;60 km/h)

Figure 3. Speed versus Fuel Cost Curve

Other VOC models suit for Australia condition are mentioned by Naude C et al. (2015).Though those models has equations which can be directly used, they did not give the detailed explanation of each cost components.

Table 2.Vehicle Operating Cost at different speed per Kilometer by vehicle type ($)

2.4．Value of Time (VOT)

VOT is a monetary value of time converted by a time cost, which is mostly determined by the hourly wages of truck drivers (Belenky, 2011). VOT is an important factor affecting drivers’ route choice behavior. Different types of vehicles have different sensitivity to VOT. Detailed data was obtained from the Freight Vehicle Calculator for the most recent hourly wages of drivers are presented in Table 3.

Table 3.Hourly wage for drivers by vehicle type ($)

2.5．Environmental Cost (EC)

Environmental and social costs are both transport externalities. Unlike VOC and VOT, these costs are not fully borne by road users, which means that travelers usually will not take these costs into account when they make route choice decisions. That is the reason they are called “external cost”. (MOVE D G, 2014)

Environmental Costs includes emission costs and noise costs. Referring to the recent document, MOVE (2014) have given a detailed introduction of the model to estimate the monetary value of EC, and presented figures from Europe. The European values have been compared with Australian (NGTSM, 2015) values and found to be same. It has proposed that each cost component consists of several elements. (For example, there are Health costs, Years of human life lost, Crop losses, Building damages, Costs for nature and biosphere caused by Air pollution). In next step, models for each element are developed and then finally converted to a monetary value. The amount of emissions discharged by different vehicle types is affected by fuel consumption and vehicle’s age (NGTSM, 2015).The more years vehicles have been used the more emissions they would discharge, especially for NO2 and PM2.5. The analysis presented below illustrates this.

Emission costs vary by year of vehicle manufacture for a 5-axle articulated at speed of 40km/h on highway is presented below in Table 4. It clearly shows that vehicle’s age affect emission/environmental cost, due to that old vehicles discharge more NOx and PM2.5. Information about the proportion of vehicles in different manufacture period can found in ABS surveys.

Table 4.Emission Cost by vehicle age ($)

Source: (NGTSM, 2015)

Noise has annoyance impacts and can damage personal health. For noise costs, Jain and Parida (2001) proposed models to detect noise levels in rural highways when road characteristics are given. Noise emissions levels can be affected by type of infrastructure, type and condition of vehicle and the vehicle’s speed. Specific noise costs per kilometer data was obtained from European studies (MOVE, 2014).

Table 5.Environmental Cost on Highway and Freeway per Kilometer by vehicle type ($)

2.6．Social Cost (SC)

Social cost considered were crash costs, congestion costs and infrastructure costs. Crash costs include the cost of injury and death, medical services, emergency services and related administration, as well as damage to vehicles and property. Specification of the number of fatal crashes and heavy/slight injuries is very important (BITRE, 2006). Rates obtained from Vicroads were used. Then average cost per crash by crashes types are presented by NGTSM so that the cost per kilometers can be evaluated.

Congestion costs are not only vehicle-related costs and labor cost but also the cost of negative effects on society, which include boosting pressure on residents, disturbing people’s daily plans, making time plans unreliable and increasing traffic accidents. Different congestion levels cause different costs. Speeds were used to define congestion levels (MOVE, 2014). Making an assumption that on freeways, free-flow condition is when average speed are over 100 km/h while a severe congestion linked with an average speed that lower than 30km/h. On Highway’s, these two figures are 80km/h and 20km/h, respectively. We used an exponential curve to estimate congestion cost at various speeds.

Infrastructure costs contain the maintenance of road, construction cost and administration cost. The data is mainly from the latest European report (MOVE D G, 2014).

Table 6.Social Cost on Highway and Freeway per Kilometer by vehicle type ($/km)

From the Table 6, on freeway (highway) free flow means that the speed is more than 100km/h (80km/h) while near capacity means at speeds lower than 30 km/h (20km/h).

3. Analysis and Discussion

In order to elaborate the optimum toll charge price a case study was used in this analysis. An arbitrary O-D pair was selected and alternative route options were found with the help of Google maps as shown below in Figure 4. Route one is via Eastlink and alternative route is via (Name) Highway in City of Melbourne. It is important to note that this is just an example considered for explanation purpose and conditions may vary based on many parameters such as O-D pair, time, days, season, etc.

Eastlink has varying toll charges based on vehicle category and for this road segment, for selected case, the toll charge was found to be $5.70 for light commercial vehicles (LCV) and $9.42 for heavy commercial vehicles (HCV).The two alternative routes selected for this analysis are, one in freeway (toll road) and the other using Highway which has a similar distance 17.1 km and 17.9 km, respectively.

Figure 4.Alternative routes selected for analysis

The main objective of this analysis is to compare/contrast route selection mechanism of freight vehicles mainly based on their cost factors and to look at the overall condition of traffic movement (City logistic concept) when such decisions are made. Since traffic congestions are un-avoidable, especially on highways, travel times were recorded for different times during a typical week day to see the variation in peak and off-peak times and depicted in Table 2. In Melbourne, 7:00 am and 5:30 pm times could be considered as morning and evening peak times.

Table 7.Travel Times on Freeway and Highway in different time of a day

It is a well-known fact in the transport industry that road users choose their route based on direct costs, which is an addition of vehicle operating costs and value of time. In other words, the route decision will be made purely based on cost factors including time saving component and users are not concerned about externalities they produce, especially by freight vehicles even though they are of a higher magnitude compared to passenger vehicles. Considering this, transport planners set their optimal toll levels based on the operation cost difference of two alternative facilities. On the other hand there is an argument that toll charges should cover the externalities produced by each trip since direct costs are beared by the road user but the externalities are borne by society as a whole (Holguin-Veras J et at.,2006). Therefore, in this paper will discuss all these scenarios based on a case study to evaluate impacts and optimal way of setting toll charges for freeways. Since different travel times are observed at different times of the day and this can affect the analysis, only an average conditions were used for detail analysis from this point. Therefore, the travel time selected will be 13 min for freeway conditions and 27 min for highway conditions. Based on that the table 8 was developed, which shows vehicle operating costs and value of time on freeway and highway with respected to selected situation based on the equations explained in the previous sections. The freeway alternative was considered to be operating under free-flow conditions, since the average speed observed was 79 km/hr and highway alternative is operating under stop-start condition where the average speed was found to be 40 km/hr. The cost components are calculated accordingly.

The toll charge is a sensitive price. Users try to minimise their travel cost (direct) under rational conditions but authorities should try to minimise the overall cost of transportation, which includes externalities, by means of tools and policies. For freight vehicles, since they operate in business environment they have their own models to calculate cost and optimise their transport operation, but no one is concerned to calculate the overall cost and to find out ways and which how this could be minimised. Usually the toll charges are set in agreement with the government. The next section focuses on elaborating the sensitivity of toll charges and its impacts.

The following curve depicts the typical decision making probabilities of road users with respect to use toll roads in comparison with toll charges and the differences in direct costs. Operator benefit means that how much money operators could save when using toll road.

剩余内容已隐藏，请支付后下载全文，论文总字数：56338字