U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
| REPORT |
| This report is an archived publication and may contain dated technical, contact, and link information |
|
 |
| Publication Number: FHWA-HRT-13-026 Date: March 2014 |
Publication Number: FHWA-HRT-13-026 Date: March 2014 |
To provide analysts and modeling managers with a frame of reference on the level of effort required to complete a microsimulation analysis, a sample of completed projects was assembled and analyzed. This chapter presents a summary of four different microsimulation case studies that have been successfully completed. The cases range in size (small, medium, and large) and were completed with different simulation software packages.
For the purposes of this report, the model sizes were categorized into small, medium, and large models, which are defined as follows:
The selected case studies were successfully completed projects that followed the seven-step modeling process as outlined in the FHWA report, Traffic Analysis Toolbox Volume III: Guidelines for Applying Traffic Microsimulation Modeling Software.(1) The labor-hour estimates should be used as a point of reference and not as an absolute number to apply to projects. Seemingly similar projects can require different levels of effort for many reasons including the following:
A summary of each of case study includes the following information:
The level of effort demonstrated in these case studies does not include the effort required to prepare traffic forecasts. Traffic forecasts were either prepared under a separate task or provided externally.
In all the cases provided, the number of model runs required was established externally by the performing agency. None of the projects conducted a statistical analysis of the data as a basis for the tolerance levels to be used for determining the minimum number of model runs required. Chapter 6 of this report provides a technique for conducting this exercise.
Small Model: I 35 in Forest Lake, MN
Project Description
Improvements to the interchange of County State Aid Highway 2 (CSAH 2) with I-35 were proposed by Washington County in Minnesota. As part of the requirements for approving the access modification, a microsimulation model of the study area was prepared. The purpose of the modeling effort was to determine the impact that different interchange concepts would have on I‑35 and to use this information to select a preferred alternative. Figure 1 shows a map of the study area.
Figure 1. Illustration. I-35/CSAH 2 study area.
Model Size
The model consists of approximately 18 mi of freeway mainline, two local access interchanges, and one regional access interchange. The temporal extents of the model were 3 h in the morning and afternoon peaks, and the traffic flow data were broken down into 15-min intervals. The model has 154 nodes and 174 links.
The model included I‑35 (north-south interstate) with three crossing arterials (interchanges) with ramp terminal intersections at each location (six intersections). There were no parallel arterials adjacent to I‑35 included in the microsimulation model.
The traffic analysis for this project was one component of the overall planning, design, and construction of the CSAH 2 interchange. The total level of effort required to complete the analysis is summarized in table 1. The data collection involved a combination of manual counts and an assembly of available counts from the Minnesota Department of Transportation (MnDOT) Traffic Management Center.
Table 1. Small model I 35 level-of-effort summary.
| Task Number | Task | Hours |
| 1 | Data collection | 40 |
| 2 | Simulation model development and calibration | 140 |
| 3 | Design year traffic analysis (four alternatives) | 200 |
| 4 | Opening year analysis | 80 |
| 5 | Documentation | 40 |
| 6 | Project management | 40 |
| Total | 540 | |
Medium Model: State Trunk Highway 100 in Saint Louis Park, MN
Project Description
The purpose of this project was to test proposed improvements on State Trunk Highway 100 (TH 100) between Excelsior Boulevard and Cedar Lake Road. The project was not the subject of a formal interstate access request for FHWA; however, due to the complexity of the interchanges and the levels of congestion (i.e., in excess of 8 h per day), MnDOT required that a microsimulation analysis be conducted. TH 100 was the first freeway built in Minnesota and was originally constructed in the 1930s as a four-lane freeway with interchange access every 0.5 mi. Over time, TH 100 was increased to a six-lane freeway north and south of the study area. The modeled area and the project limits are illustrated in figure 2.
The proposed improvements were intended to provide a full standard six-lane freeway with interchange access consolidated to approximately 1 mi. The access was consolidated through a series of a collector distributor roads and frontage roads. The original project scope entailed examining one design alternative; however, the analysis revealed there were operational challenges with the proposed design, and the alternatives analysis task was expanded to include the development of new design concepts.
Figure 2. Illustration. TH 100 study area.
Model Size
The project area simulation limit was 13 mi. The I‑394 freeway is immediately to the north of the project area and has an impact on operations within the project area. A portion of I‑394 (east-west) was included to account for these traffic impacts. The model did not include parallel facilities, and there were eight signalized intersections included in the model. The model has 659 nodes and 646 links.
Level of Effort
The level effort involved directly supporting MnDOT staff who developed the preliminary plans, conducted public involvement meetings, and sought local approval of the project. Table 2 provides a summary of the level of effort expended to complete the analysis. The data collection portion of the project consisted of downloading traffic management center data, compiling the data, and smoothing the field observations. Intersection turning movement counts had previously been conducted by MnDOT; these costs are not reflected in the level of effort.
Table 2. Medium model TH 100 level-of-effort summary.
| Task Number | Task | Hours |
| 1 | Data collection | 262 |
| 2 | Building of base model | 339 |
| 3 | Base model calibration | 258 |
| 4 | Design year traffic analysis | 871 |
| 5 | Opening year traffic analysis | 185 |
| 6 | Documentation | 62 |
| 7 | Project management | 84 |
| Total | 2,061 | |
Medium Model: I 5 in Tacoma, WA
Project Description
The purpose of this project and analysis was to obtain approval from FHWA to modify interchanges along I‑5 in Tacoma, WA, to accommodate a new high-occupancy vehicle (HOV) lane. The modeled area and the project limits are illustrated in figure 3. The project had been studied previously, and the level of effort that is described in this guidebook represents a comprehensive update of the previous effort. The alternatives analysis considered only the alternative that had been previously selected.
©Google Earth® 2012
Figure 3. Illustration. I-5 Tacoma study area.(2)
Model Size
The project area spanned 10 mi of I‑5, and the simulation limits included a total of 10 mi. The project did not include parallel arterials, and there were 27 intersections included in the model. The model has 164 nodes and 203 links.
Level of Effort
The level of effort to conduct the I‑5 project in Tacoma, WA, is summarized in table 3. The level of effort for this project was streamlined due to previous work efforts related to the project. Washington State Department of Transportation (WSDOT) conducted the traffic data collection.
Table 3. Medium model I 5 Tacoma level-of-effort summary.
| Task Number | Task | Hours |
| 1 | Data collection* | 80 |
| 2 | Building of base model | 160 |
| 3 | Base model calibration | 400 |
| 4 | Design year traffic analysis | 320 |
| 5 | Opening year traffic analysis | 320 |
| 6 | Documentation | 240 |
| 7 | Project management | 80 |
| Total | 1,600 | |
| *Traffic count data were supplied by WSDOT. | ||
Large Model: I 5 in San Diego, CA
Project Description
The purpose of this project was to prepare a Corridor System Management Plan (CSMP) for the California Department of Transportation (Caltrans) district 11 and the San Diego Association of Governments. The modeled area and the project limits are illustrated in figure 4.
Figure 4. Illustration. I-5 San Diego study area.
Model Size
The project includes 30 mi of freeway (27 mi on I‑5 and 3 mi on I‑805), 5 system-to-system freeway interchanges, and 26 service interchanges. The model has 659 nodes and 796 links.
Level of Effort
Table 4 provides a level-of-effort summary for the I‑5 San Diego case study.
Table 4. Large model I-5 San Diego level-of-effort summary.
| Task Number | Task | Hours |
| 1 | Select analysis tool | 60 |
| 2 | Collect data | 2,000 |
| 3 | Develop and calibrate baseline model | 2,500 |
| 4 | Develop future baseline model | 860 |
| 5 | Analyze alternatives | 1,660 |
| 6 | Finalize documentation | 1,200 |
| 7 | Develop and conduct presentations | 650 |
| 8 | Develop work plan and project management | 1,150 |
| Total | 10,080 | |
The case studies are reflective of typical traffic analysis for geometric and operation design projects. This analysis is customarily performed during project development by State transportation departments and reviewed by USDOT staff for interstate access and other requirements. The case study projects were not intended to address traffic analysis for planning studies, regional analysis projects, different operational conditions (i.e., impacts of weather events, high travel demand, and incidents), or multimodal projects where mode shift may occur.
Each of the case study projects is unique in terms of alternatives being considered. A common factor among all the projects is the building of a microsimulation model following the basic seven-step procedures described in Traffic Analysis Tool Box Volume III: Applying Microsimulation Model Software.(1) All of the case studies ran the simulation models multiple times with different random number seeds and reported the averaged model results.
There are some similarities and differences between the four sample projects. The similarities include base model development and calibration and indicate that the modeling efforts are scalable to the network size (e.g., number of hours per interchange). The other tasks that are impacted by local and project specific factors include the number and type of alternatives being tested. Table 5 provides a composite comparison of the four case studies by major task category.
Table 5. Comparison of level of effort of sample projects.
| Task | Model Size and Labor Hours | |||
| Small | Medium | Medium | Large | |
| Data collection | 40 | 262 | 80 | 2,000 |
| Base model development and calibration | 140 | 597 | 560 | 2,500 |
| Alternatives analysis | 280 | 1,056 | 640 | 2,520 |
| Documentation/presentations | 40 | 62 | 240 | 1,850 |
| Project management | 40 | 84 | 80 | 1,210 |
| Total | 540 | 2,061 | 1,600 | 10,080 |
Total Level of Effort Compared to Base Model and Calibration
Figure 5 provides a comparison of the total level of effort and the amount of effort expended specifically for building a base model and the calibration process. The only apparent trend is a consistent order of magnitude regarding the model building and calibration effort with the model size. Larger models require a disproportionately greater level of effort compared to smaller projects. This is attributed to the increased complexity of a larger simulation model as it relates to modeling travel demand and traffic operations in a larger geographic area and extensive project reporting, alternatives analysis, and presentation time expended.
Figure 5. Graph. Total level of effort comparison.
Percentage Breakdown of Task Allocation
A further breakdown of the percent allocation of time was reviewed. Figure 6 shows the composite percentage of hours by task for all four case study projects. Figure 7 shows a composite percentage of hours by task for the three small/medium case study projects.
Figure 6. Graph. Composite effort percentage by major tasks for all four case studies.
Figure 7. Graph. Composite effort percentage by major tasks for small- and medium-sized models only.
Comparison Conclusions
Larger models require a disproportionately greater level of effort compared to smaller projects. This is attributed to the increased complexity of a larger simulation model as it relates to modeling travel demand and traffic operations in a larger geographic area and extensive project reporting, alternatives analysis, and presentation time expended.
Roughly a third of the level of effort on the projects was allocated to the base model development and calibration. Model calibration is a critical activity in microsimulation and can drive the project level of effort to high levels. The experience level of project managers, analysts, and reviewers, as well as a robust quality assurance process can help control the risk of budget and time overruns.
Depending on the availability of data, the amount of time allocated to data collection varied from 8 to 16 percent of the total level of effort across the four case studies. As some of these locations had considerable existing instrumentation and resulting data available for their interstate system, these values may be lower than what other locations may experience. Anecdotal information from other non-instrumented regions indicates that up to 30 percent of the project effort may be devoted to data collection in the absence of archived data.
