The identification and selection of performance measures play an important role in any decision making process. Additionally, millions of dollars are spent on appropriate planning and identification of prospective projects for improvements. As a result, current practitioners spend a lot of time and money in prioritizing their limited resources. This research proposes two tasks: 1) estimation of performance measures using a simulation based on dynamic traffic assignment model, and 2) development of a methodology to evaluate multiple projects based on benefit-cost analysis. The model, DynusT, is used for the Las Vegas roadway network during the morning peak time period. A comparative analysis of the results from proposed methodology with existing California Benefit-Cost (Cal-B/C) models is presented. The results indicate that the new methodology provides an accurate benefit-cost ratio of the projects. In addition, it signifies that the existing Cal-B/C models underestimate the benefits associated with the prospective project improvements. The major contribution of this research is the simultaneous estimation of the performance measures and development of a methodology to evaluate multiple projects. This is helpful to decision makers to rank and prioritize future projects in a cost-effective manner. Planning and operational policies for the transportation systems can be developed based on the gained insights from this study.
The identification and selection of performance measures play an important role in any decision making pro- cess. This helps the policy makers to allocate appropriate resources for prospective future improvements and evaluate projects. A myriad of literature is available that captures multiple performance measures within the Transportation, Activity and Environmental systems [
There are primarily two type of models to assess effects on traffic safety; accident-risk-based models (ARBM), and accident prediction models (APM). ARBMs are descriptive models based on traffic accident and exposure data whereas APMs are based on available data to quantify the relationship between accidents and traffic characteristics (speed or flow). The ARBM assumes that the individual probability of being involved in a crash increases linearly with exposure. Lord [
Researchers have used simulation models or Travel Demand Models (TDM) to estimate emissions and fuel consumption. There are basically two types of emission models―average-speed based and instantaneous-speed based. Ahn and Rakha [
Considering the level of resolution used to model network traffic flows, modeling approaches can be categorized as macroscopic, microscopic, or mesoscopic. Normally, macroscopic approaches involve static traffic assignment models that enable the estimation of flow patterns on a regional scale but without any temporal resolution. These types of models use macroscopic traffic flow relationships to determine link travel times based on link flows. The TDM aggregates the origin-destination (OD) matrices across all modes before the traffic assignment step. As a result, the model cannot differentiate between truck and car assignments. Hence, the TDM cannot be directly used to conduct the desired analysis. In addition, the implementation of a multiclass assignment using a TDM framework requires addressing algorithmic and computational issues. The primary difference between single class and multiclass models is that the travel cost functions for the latter are non-symmetric and non-separable, hence convex optimization techniques are not applicable [
In contrast, microscopic models enable the explicit modeling of individual vehicles as well as temporal variations in traffic flow in the order of 0.1 to 1.0 seconds. In addition, they illustrate detailed traffic characteristics, such as lane changing behavior, acceleration/deceleration, and queuing related phenomena like spillback/spil- lover. However, this type of modeling requires a substantial amount of computational time and data collection efforts. As a result, it is very difficult and expensive to develop them for large-scale systems.
To overcome some of these limitations, many emerging planning strategies such as congestion pricing and the operational deployment of information provision services require modeling approaches that enable a greater level of detail than macroscopic models and with a much larger geographical scope than microscopic models. Mesoscopic models combine micro and macro level capabilities and incorporate many time-dependent traffic flow characteristics, such as spillback/spillover on a regional-level scale. The DTA models load individual vehicles into the network and solve a traffic assignment problem considering the operational characteristics of vehicles. Hence, this study requires a DTA capability that considers multiples classes of vehicles in terms of their routing strategies and behavior including trucks and regular passenger cars.
There are differences in calculating performance measures using static vs dynamic approaches. Kockelman et al. [
The paper is organized as follows. The simulation-assignment model, DynusT, and the estimation of performance measures are discussed in Section 2. Experiments are conducted in Section 3 to calculate the benefit-cost ratios for certain projects. Results and analysis are discussed in Section 4. Conclusions and recommendations are presented in Section 5.
This section describes the modeling and analysis approach. A simulation-based dynamic traffic assignment technique is used to estimate traffic flow related characteristics. Different models are used to estimate multiple performance measures based on the traffic flow characteristics. Section 2.1 discusses the network modeling approach and Section 2.2 discusses the procedure to estimate performance measures.
The simulation based DTA model used in this research is DynusT [
Two separate OD demand matrices were imported from the TDM, one for passenger cars and one for trucks. The Las Vegas roadway network includes a total of 1646 Traffic Analysis Zones. The morning peak-period (6 AM to 9 AM) was modeled using the corresponding three-hour demand that was distributed for every 15-minute time interval. Hence, a total of twelve demand matrices were used to dynamically load the vehicles into the network. Considering the demand profile, it was determined that aggregation of demand was feasible and convenient for computational performance. After aggregation, the number of zones was reduced from 1646 to 696 and the entire model was consistently updated to reflect zoning changes. Based on the input files, the DTA model was used to determine the average network traffic flow pattern for a morning peak-period of a weekday. To measure the difference between the model results and the real-world, calibration was performed and simulated link counts were compared to actual link counts. Several iterations of calibration were conducted until at least 85% of the link counts were within 15% error region, as specified by the Federal Highway Administration Traffic Analysis Volume Toolbox III [
This section provides a methodology to estimate the performance measures based on the output from DTA model. In addition, the monetary value (in dollars) associated with corresponding performance measure is also discussed. The inclusion of dollar value will help the decision makers in evaluating the projects for safety improvements. Also, this will benefit in allocating appropriate resources for overall system performance. The estimated performance measures include: Travel Time, Crashes, Emissions, Fuel Consumption and Vehicle Operating Costs.
Travel time for a network link is obtained directly from the DTA model. The peak hour volume is extracted based on hourly volumes. As a result, appropriate daily and yearly factors are used to convert it into annual travel time. A wage rate of $20/hour is recommended to compute the corresponding monetary costs associated with travel time.
Safety estimations are computed using the ITS Deployment Analysis Systems methodology, developed by the Intelligent Transportation Systems (ITS) Joint Program Office of the US Department of Transportation [
where,
The total number of crashes is equal to the summation over the entire network of the number of crashes in each link. Comparison between estimated and actual crashes [
where,
Emissions play a very important role in the evaluation of transportation alternatives because they are directly related to human health and the environment. Major pollutants from vehicles include carbon monoxide, volatile organic compounds, oxides of nitrogen, oxides of sulfur, carbon dioxide and particulate matter (PM10). This study uses Emission Rates (ER) in gm/mile provided by the California Air Resource Board [
where,
The emissions cost for each of the pollutants is obtained using Benefit/Cost models (Cal-B/C models) developed by the California Department of Transportation. It is assumed that the emissions cost in the Las Vegas Valley is the same as the cost in the Los Angeles/South Coast region. The monetary value of emissions (dollar/ton) in 2011 is based on the Cal-B/C models [
where,
Fuel consumption plays a vital role in the evaluation of investment of transportation projects. Fuel consumption rates (FC) (in gallons/mile), is obtained by EMFAC 2011 model. These rates are a function of link speeds that are obtained for each vehicle type using the simulation-based methodology. Fuel consumption for each link in the network is given by the Equation (5).
where,
Based on the 2011 gas rates, gas cost for autos is assumed as $3/gallon and diesel cost for trucks is assumed as $3.4/gallon. Equation (6) shows the fuel consumption costs for any link in the network.
where,
Vehicle operating costs (VOC) depends on vehicle usage. Components that constitute VOC include fuel, oils, tires, maintenance, repairs, and mileage-dependent depreciation [
where,
The performance measures for years 2012, 2013, 2020, and 2030 is obtained from post processing the DynusT output and converted to monetary values as discussed in Section 2.2. It is assumed that the growth in between the years is linear and an inflation adjusted rate is used to calculate the respective benefits. Finally, all the benefits for future years are converted to present year using discount rate of 7% and added up to obtain total benefits. Similarly, the costs (right of way, construction, maintenance etc.) associated with a particular project is identified and converted to present value using the discount rate to obtain costs. As a result, the benefit-cost ratio is identified for the project. The entire analysis is coded and converted into an Interface. This interface is modular and the user defines the analysis year. The interface is flexible and it can estimate the performance measures based on link, corridor, zone, or a network depending on the specified time interval. For multiple alternatives, a zone is selected for each alternative and then the interface is run for that particular scenario to check the differences from the base case. The interface doesn’t have the capability to generate results for comparing multiple alternatives simultaneously.
Ideally, for transportation performance management, two types of economic analysis are performed. The first systematic means of comparing highway investments is called life-cycle cost analysis (LCCA) [
This section discusses two techniques to obtain the benefit-cost ratio for projects in Las Vegas metropolitan area. The first one is the traditional Cal-B/C model [
The California Department of Transportation uses Cal-B/C to conduct investment analyses of projects proposed for the State Transportation Improvement Program, the State Highway Operations and Protection Program, and other ad hoc analyses requiring benefit-cost analysis. Cal-B/C is a spreadsheet-based tool that can prepare analyses of highway, transit, and passenger rail projects. The model uses input data defining the type, scope, and cost of projects. The model calculates life-cycle costs, net present values, benefit-cost ratios, internal rates of return, payback periods, annual benefits, and life-cycle benefits [
The benefit-cost analysis on three federally funded projects sponsored by the Nevada Department of Transportation (NDOT) was performed using Cal-B/C models. The analyses were formed from existing project reports and NDOT databases that contained project data. The benefit-cost analyses were performed using Cal-B/C with parameter and rate adjustments based on local conditions for Nevada. The following performance measures were considered in the evaluation of benefits and costs.
・ Travel Time Savings
・ Accident Reductions
・ Vehicle Operating Costs
・ Vehicle Emission Reductions
・ Pavement Roughness
・ Project Capital Costs
・ Project Operation & Maintenance Costs
These analyses all use a 20-year horizon to enable comparisons among each other. The analyses use a real discount rate of 7% as recommended by the Office of Management and Budget (OMB) Circular A-94 [
The proposed methodology uses the output obtained after running DynusT through the entire network. For analysis, a zone is selected near the proposed project.
boundaries for one of the projects in Network EXplorer for Traffic Analysis (NEXTA). NEXTA is an interface used to facilitate the preparation, post-processing, and analysis of simulation-based dynamic traffic assignment datasets. The benefit-cost methodology uses an interface as shown in
Once improvements to the project are made, DynusT is run again to obtain a new set of output. The same zone is selected again and using the previous input data, the process is repeated to estimate the total costs. Finally, the net benefit of the project is the difference in total costs before and after the improvements. The net cost is the cost involved in project improvement/construction.
For any project,
rizon with a discount rate of 7%. The x axis represents the years whereas y axis represents total travel time in billions of hours.
The comparative analysis of the results of the benefit-cost ratio obtained from the proposed model and the Cal-B/C models is shown in
The results from
Existing state of the art techniques concentrated primarily on estimation of performance measures using static approaches. However, to accurately estimate the traffic flow characteristics, dynamic models were predominately used by researchers. This research proposed a comprehensive methodology to estimate performance measures using DTA models and evaluate projects. The evaluation was done using the benefit-cost analysis techniques. Numerical experiments were conducted to evaluate three projects in Las Vegas Metropolitan area. A
Project No. | Project description | Type | Benefit-cost ratio from Cal-B/C models | Benefit-cost ratio from proposed model |
---|---|---|---|---|
1 | North 5th Street Super Arterial Phases 1C & 1D: Carey to Cheyenne | Bridge Construction | 12.60 | 13.68 |
2 | Boulder City Bypass Phase 1: Foothills Drive to US-93/US-95 Interchange | Bypass/New Interchange | 0.90 | 4.25 |
3 | US 93 Pavement Rehabilitation & Truck Climbing Lanes | Widening/Pavement Rehabilitation | 8.30 | 24.17 |
comparative analysis with the existing Cal-B/C models revealed that the proposed methodology provides an accurate benefit-cost ratio. In addition, the results also indicated that Cal-B/C models underestimate the benefits associated with the projects. The experiments showed that the proposed methodology is robust and it provides a suitable technique for decision makers to rank and prioritize projects. Planning and operational policies for the Transportation systems can be developed based on the gained insights from this study.
The major contribution of this research work is the simultaneous estimation of the performance measures and development of a methodology to evaluate multiple projects. However, there are certain limitations associated with this research. The comparative analyses presented here are for three projects, and certainly more projects could be added to provide deeper understanding of the differences among the benefit-cost ratio results. In addition, the interface presented in the proposed methodology can be improved further to generate results for comparing multiple alternatives simultaneously. Future work will attempt to address the above limitations.