National Academies Press: OpenBook
« Previous: Front Matter
Page 1
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 1
Page 2
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 2
Page 3
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 3
Page 4
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 4
Page 5
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 5
Page 6
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 6
Page 7
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 7
Page 8
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 8
Page 9
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 9
Page 10
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 10
Page 11
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 11
Page 12
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 12
Page 13
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 13
Page 14
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 14
Page 15
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 15
Page 16
Suggested Citation:"Executive Summary." National Academies of Sciences, Engineering, and Medicine. 2013. Dynamic, Integrated Model System: Jacksonville-Area Application. Washington, DC: The National Academies Press. doi: 10.17226/22482.
×
Page 16

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

1Executive Summary Purpose and Need SHRP 2 Project C10A, Partnership to Develop an Integrated, Advanced Travel Demand Model and a Fine-Grained, Time-Sensitive Network, was undertaken to develop a dynamic, integrated model and to demonstrate its performance through validation tests and policy analyses. Key goals of the SHRP 2 C10A model system development effort include providing enhanced rep- resentation of travelers’ sensitivities to price and incorporating findings from other SHRP 2 Capacity projects. Modeling Travel Travel models are used to support decision making by providing information about the impacts of transportation and land-use investments and policies, as well as demographic and economic trends. When applied properly, they can provide a consistent framework for evaluat- ing different alternative scenarios. Transportation decision makers need to have confidence that the tools they use to inform policy and investment decisions, including travel demand forecasting models, produce reasonable results that are appropriately sensitive to the questions at hand. Most travel models comprise a set of components that address different aspects of traveler choices. The four steps of travel models involve determining (1) the number and purpose of trips to be made, (2) the origins and destinations of those trips, (3) the travel mode (such as driving alone or riding transit), and (4) the specific network routes used. These steps can be broadly grouped into demand and supply categories, with the first three (generation, distribution, and mode choice) describing the demand components and the last one (assignment) describing the supply components. Recent methodological advances have occurred with both model demand components and model supply components. These advances provide the opportunity to develop more robust travel models for use in transportation decision making. Activity-Based Demand Models On the demand side, metropolitan planning organizations (MPOs) have increasingly adopted activity-based models. Activity-based travel demand models produce estimates of daily activity patterns including tour and trip generation, destination choice, mode choice, and time-of-day choice. A tour is a chain of trips that begin and end at home or work; it is essential for represent- ing the interrelationships between activities undertaken by travelers. Daily activity pattern mod- els consider the coordinated aspects of travel made by an individual across the entire day, as well as activities potentially coordinated across individuals within a household. In addition, these

2models typically incorporate accessibility measures that allow changes in network performance to influence demand generation. Agencies generally develop and apply activity-based models to include sensitivities to policies that may be challenging to represent in a traditional, trip-based models. For example, the effects of pricing policies on demand generation, destination, and mode choices can be better captured using activity-based models than trip-based methods. A number of features distinguish activity- based approaches to modeling demand from traditional trip-based approaches. These features include the following: • Activity-based models represent travel demand in a more intuitive manner than tradi- tional, trip-based demand models because they simulate individual and household travel choices. For example, a traveler decides whether to make a tour or stop to participate in an activity where that activity will take place (such as whether to work at home or journey to work), and when and how to get there, in an intuitive way that captures opportunities and constraints. • Activity-based models provide more consistency, and potentially more detail, across all dimen- sions of travel behavior, especially space and time. In turn, that consistency results in more realistic representations of transportation system performance by the network supply model. Significantly, activity-based models do not include non–home-based trips. Those trips fre- quently make up a large portion of the demand in traditional trip-based models, but trip-based models cannot include potentially relevant information such as prior trip-mode choices and traveler income. • Activity-based models include significantly more detail on traveler attributes and spatial and temporal constraints, which provides better estimates of the transportation impacts of a given alternative scenario. For example, activity-based models can assign person-specific and purpose-specific values of time to different individuals, which is important for modeling pric- ing alternatives. Such detailed market segmentation is possible because of the disaggregate nature of most activity-based model implementations and is often intractable in the context of aggregate trip-based travel models. • Activity-based models produce a wider range of performance measures, with greater detail. Perhaps most significantly for the C10A project, activity-based demand models explicitly include a detailed representation of time-of-day using temporal units of half-hours and minutes rather than broad multihour time periods. This temporal resolution facilitates the incorporation of changes in network performance by time-of-day that are produced by the dynamic supply model; it also provides an explicit method for reflecting availability constraints (such as time- window accounting), which produces activity patterns that are logically consistent in both time and space. Dynamic Traffic Assignment Models On the supply side, metropolitan planning organizations are also increasingly adopting dynamic traffic assignment (DTA) approaches. Traffic assignment is the fourth and final step of the tra- ditional four-step planning process. Until the past decade, virtually all travel models incorpo- rated static traffic assignment methods, which produce estimates of travel times, costs, and volumes across relatively broad time periods. However, the analysis and management of trans- portation network performance require information about time-varying network times, costs, and flows, and transportation policies increasingly incorporate time-varying assumptions. Static-based assignment approaches cannot represent time-varying flows and congestion or the impacts on travel times and costs with sufficient detail. In contrast, dynamic network models do have the ability to represent time-varying network time and costs; in addition, they can provide

3more information on network performance by detailed time of day, which can be used as input into travel model demand components. Features that distinguish dynamic network methods from static network approaches include the following: • DTA models incorporate more complete representations of transportation network attributes and configurations, including better representations of intersection controls such as signal synchronization and other advanced network control schemes. • DTA models use more realistic flow models to propagate traffic on links, rather than using sim- plified volume-delay functions, which may produce unrealistic estimates of network times and volumes. • DTA models provide more detailed estimates of network system performance, which is essen- tial for accurately evaluating the impacts of different transportation policy, systems manage- ment, and funding alternatives. Integrated Models Transportation policy and investment questions have become increasingly complex. At the same time, existing models have been made more behaviorally descriptive, and new models have been developed. Separate models that had been viewed as independent are now often viewed as interdependent. The purpose of the C10A project was to make operational a dynamic, integrated model and to demonstrate its performance through validation tests and policy analyses. An integrated model system is essential because most current travel models are not sufficiently sensitive to the dynamic interplay between travel behavior and network condi- tions; they are unable to reasonably represent the effects of transportation policies, such as variable road pricing and travel demand management strategies. The availability and capabili- ties of activity-based demand models and dynamic network supply models provide the oppor- tunity to address the shortcomings of current tools and provide decision makers with more complete information. Project Objectives As stated, the primary objective of the C10A project was to make operational a dynamic, integrated model—an integrated, advanced travel demand model with a fine-grained, time-dependent net- work. The model’s performance would then be demonstrated through validation tests and policy analyses. Secondary project goals included producing a transferrable process and sample data for use in other regions, demonstrating an effective interface with the Environmental Protection Agency’s (EPA’s) motor vehicle emission simulator (MOVES) model, addressing travel time reli- ability in travel models, and incorporating knowledge from other SHRP 2 efforts. These include Project C04, Improving Our Understanding of How Highway Congestion and Pricing Affect Travel Demand (pricing), and Project C05, Understanding the Contribution of Operations, Tech- nology, and Design to Meeting Highway Capacity Needs (operations). This report describes the tools incorporated into the integrated model system, the data required to implement the tools, modifications to the tools that were necessary to achieve the integration, and results of sensitivity tests of the integrated model system. The C10A project team envisioned implementing the project in a region with limited choices of nonhighway modes; as such, the dynamic, integrated model represents behavioral changes in response to roadway conditions. To meet this objective, the model system was designed to cap- ture changes in demand, such as time-of-day choice (i.e., peak spreading), and route choice in response to capacity and operational improvements, such as signal coordination, freeway man- agement, variable tolls, and capacity improvements. While the primary project objective called for the development of a dynamic, integrated model with advanced policy analysis capabilities, the project team also noted that advanced practitioners

4have to be able to implement the model system in other regions without excessive costs or undue complexity. The resulting model system has the following features: • The model is scalable. While the model system implemented for C10A can exploit distributed computing to reduce model system runtimes, it does not require a large hardware cluster or complex computing environment. • The model is relatively easy to implement and maintain. Although the model system is inher- ently complex because of its advanced capabilities, it can be easily and flexibly configured to operate with different levels of temporal and spatial detail and in different computing environments. • The model system does not require a multiyear, multi-million-dollar implementation and maintenance effort. The model system was implemented in two regions and subjected to a set of initial calibration and sensitivity tests in approximately 1 year. This report documents the implementation of the model system in both Burlington, Vermont, and Jacksonville, Florida; the calibration and validation of the model system; and the application of the model system to a set of initial sensitivity tests. Model System Components The proposed model system comprises three primary components: DaySim, the Transportation Analysis and Simulation System (TRANSIMS) Router and Microsimulator, and MOVES. DaySim is a travel demand forecast model that predicts household and personal travel choices at a parcel level on a minute-by-minute basis. The TRANSIMS Router and Microsimulator are dynamic traf- fic assignment and network simulation software that can perform regional traffic microsimulation on a second-by-second basis. MOVES is the EPA’s latest software for estimating emissions and air- quality impacts. The C10A integrated model links these three components in an equilibrated model system that provides enhanced policy sensitivities at significantly higher levels of spatial and tem- poral resolution than are found in traditional regional travel demand forecasting systems. DaySim The travel demand model used for this project is coded in a software framework called DaySim. DaySim is one of the two main families of activity-based model systems now being used by MPOs in the United States. DaySim was initially implemented in Sacramento, California, and has been enhanced to interface effectively with the TRANSIMS tools. DaySim simulates 24-hour itineraries for individuals with spatial resolution as fine as indi- vidual parcels and temporal resolution as fine as single minutes, so it can generate outputs at the level of resolution required for input into dynamic traffic simulation. DaySim’s predictions in all dimensions (activity and travel generation, tours and trip-chaining, destinations, modes, and timing) are sensitive to travel times and costs that vary by mode, origin–destination (O-D) path, and time of day; thus it can, in turn, effectively use as inputs the improved network travel costs and times output from a dynamic traffic simulator. DaySim is structured as a series of hierarchical or nested choice models. The general hierarchy places the long-term models (such as auto avail- ability) at the top of the choice hierarchy and the short-term models (such as trip-mode and time- of-day choice) at successively lower levels in the hierarchy. More details of the DaySim structure and capabilities and a description of the DaySim-TRANSIMS linkage are provided in Chapter 1. TRANSIMS TRANSIMS network and travel assignment processes are used to represent the performance of the transportation networks in the integrated model system. TRANSIMS assigns a sequence

5of trips or tours for individual household persons between specific activity locations (smaller than travel analysis zones but larger than individual parcels) to roadways, walkways, and tran- sit modes on a second-by-second basis for a full travel day. The TRANSIMS networks include detailed information regarding the operational characteristics of the transportation facilities that may vary by time of day and by vehicle or traveler type. This information includes the number of lanes; the lane-use restrictions; the traffic controls, signal timing, and phasing plans; turning restrictions; and tolls. TRANSIMS implements a dynamic user equilibrium network assignment for trip and activity files that defines the demand by detailed time of day. The primary demand input to TRANSIMS is an activity file produced by DaySim that contains information on each individual’s activity loca- tions, timing, and mode of travel. In addition, trip list files are used to represent non–household- related travel such as trucks, external trips, and other commercial travel in the network demand. TRANSIMS tools also generate zone-to-zone network impedance measures by detailed time of day for use in subsequent DaySim demand simulations. A description of this TRANSIMS-DaySim linkage can be found in Chapter 1. MOVES The MOVES software was developed by the EPA to provide estimates of emissions and green- house gases. MOVES uses detailed information about the distribution of vehicle miles traveled (VMT) by source type, facility type, area type, time of day, day of week, and 5-mph-average-speed bins to calculate emissions for an array of pollutants. In addition to travel data, MOVES uses information about fleet-age and fuel-type distributions, inspection maintenance programs, and monthly temperatures and humidity for each county in the analysis area. These are used to cal- culate county-based emissions inventories or custom domains that combine counties into aggre- gate estimates. TRANSIMS tools have been developed to interface with MOVES to support the generation of these estimates. A detailed description of the TRANSIMS-MOVES linkage is pro- vided in Chapter 1. regional Implementations As part of the C10A project, the DaySim-TRANSIMS-MOVES model system was implemented in two regions: Burlington, Vermont, and Jacksonville, Florida. Small-Scale Regional Test Bed The integrated model system was first implemented in Burlington, Vermont. The purpose of this implementation was to establish a test bed for developing and refining model system capabilities and configurations. The Burlington modeling area comprises a single county (Chittenden) of approximately 620 mi2, and was home to 55,000 households in the base year of 2005 (Figure ES.1). These households generate approximately 525,000 daily person trips. From a development perspective, the primary advantage of implementing the model system in a smaller region is that it allows researchers to more rapidly test alternative model configu- rations and to debug model processes because the model system runtimes are shorter. Shorter runtimes are associated with both the DaySim demand component of the model system and the TRANSIMS supply component. In the DaySim model, runtimes relate directly to the amount of demand, so regions with a smaller population can be simulated more quickly. However, the DaySim demand model is not the primary performance bottleneck in the model system. The overall model system runtimes are primarily driven by the performance of the TRANSIMS Router and Microsimulator network assignment tools. Like DaySim runtimes, TRANSIMS runtimes are related to the amount of demand being simulated, but they are also significantly influenced by the level of transportation

6network detail—specifically the number of links in the network. In the Burlington implementa- tion, the TRANSIMS network is relatively coarse, and the small modeling area limits the number of network links. However, because the levels of congestion in the region are relatively low, this smaller region cannot support the full range of model system sensitivity testing required by the C10A project and the model system’s responses to a number of policies and improvements are limited. Large-Scale Regional Demonstration Subsequent to the initial model implementation in the Burlington region, the integrated model system was implemented in Jacksonville, Florida. The purpose of the second implementation was to provide a more robust and challenging context for testing the model system capabilities. The Jacksonville region comprises four counties in northeast Florida covering 3,100 mi2 (Figure ES.2). The regional population includes more than 525,000 households and 1.25 million people and generates more than 4 million daily person trips. From the perspective of model application and sensitivity testing, the purpose of imple- menting the model system in the larger region was to subject the model system to a broader and more rigorous set of policy sensitivity tests. The region’s higher levels of network conges- tion made this possible. The primary disadvantage of using this larger model region is that the additional demand and network detail result in significantly longer model system runtimes, primarily attributable to the TRANSIMS Microsimulator. Use of this larger area for model development would have resulted in a longer model development phase. The longer runtimes associated with the Jacksonville integrated model implementation also necessitated the devel- opment and testing of a number of alternative application modes, as described in the follow- ing section. Application Modes The primary driving force behind the SHRP 2 C10A project is the need to address transporta- tion policies being considered by MPOs around the country that are not adequately addressed by current state-of-the-practice travel-forecasting models. The integrated, time-sensitive Figure ES.1. Burlington model area.

7model developed for this project seeks to address the following broad categories of policies and strategies: • Pricing policies; • Capacity enhancements; • Transportation system management (TSM) (and operations) improvements; • Travel demand management policies; and • Greenhouse gas reduction strategies. Each of the three primary components of the integrated model system—DaySim, TRANSIMS, and MOVES—provides unique capabilities and can be flexibly configured to address the differ- ent analysis needs associated with the different policies and strategies. The project team devel- oped different methods of combining and linking the model system components in application as a result of practical experience in working with and testing the model system. Specifically, some policies or improvements (such as roadway pricing) require regional-scale analysis, but regional-scale microsimulation can result in excessively long runtimes while adding little policy- specific sensitivity. Conversely, smaller-scale policies or improvements (such as signal coordina- tion in a corridor) may not be expected to affect overall regional travel patterns, but they may require the local sensitivities of a traffic microsimulation model. To balance policy analysis needs against practical runtime considerations, the project team devel- oped a set of model system application modes: planning, operations, and planning + operations. Table ES.1 illustrates some typical Jacksonville model system runtimes for these application modes when implemented and distributed on the Transportation Research and Analysis Com- puting Center (TRACC) computing cluster at Argonne National Laboratory. Note that runtimes are highly dependent on the particular hardware being used, the specific versions of the software tools employed (which are updated frequently), and the level of convergence required for a par- ticular analysis. As computing power increases, runtimes are expected to decrease. The following sections describe the configuration of these application modes and identify the types of policies or improvements that each might most effectively test. Figure ES.2. Jacksonville model area.

8Planning Mode The planning application mode can be used when the analysis needs are expected to result in regional-scale changes in overall levels of travel demand or changes in regional travelers’ destina- tion, mode, or time-of-day choices but are not expected to be significantly affected by local-scale traffic dynamics. The planning mode integrates the DaySim demand model with the TRANSIMS supply model in an iterative feedback loop in which DaySim outputs estimates of travel demand at the level of individual minutes for routing within the TRANSIMS Router. Temporally detailed network impedance skims based on these Router assignments are then generated and fed back as input to DaySim. A full-scale regional TRANSIMS microsimulation may be optionally run as a post process after the integrated DaySim-TRANSIMS Router application. Planning Mode Configuration The distinguishing feature of the planning application mode is that only the TRANSIMS Router is used in an integrated way with DaySim; the TRANSIMS Microsimulator is available for post- processing. The TRANSIMS Router operates at detailed temporal resolutions (such as 5 min or 15 min) and incorporates important features (such as time-dependent, shortest path building), but it uses traditional volume-delay functions (VDFs) to convert assigned volumes into 5-min or 15-min measures of link delay. These VDFs are inferior to Microsimulator-based delays in which the travel times and costs experienced by individual travelers are used to directly generate the times and delays used in path building. The Router also lacks some key functionalities of the Microsimulator, such as using actual signal timings to estimate delays at intersections instead of relying on fixed delays derived from prior assignment iterations. A critical advantage of using only the Router in the planning mode is that it runs relatively quickly even at a regional scale because it can be partitioned across multiple processing cores. By incorporating the Microsimulator as a postprocess, the impact of strategies and policies on regional and local traffic dynamics can also be assessed, albeit not in an integrated way. Figure ES.3 illustrates the configuration of the model system components in the planning mode, including network impedances based on VDFs and measures of effectiveness (MOEs) for MOVES. Table ES.1. Application Mode Runtimes Mode Planning Operations Planning + Operations Runtime for Operations DaySim demand estimation (hours) 4.0 4.0 4.0 Assignment iteration (hours) 0.5 5.0 5.0 Convergence checking (hours) 1.0 1.0 1.0 Skimming procedures (hours) 1.0 0.0 1.0 Total (hours) 6.5 10.0 11.0 Iterations Assignment 40 40 40 System 3 1 3 Total 120 40 120 Total System Runtime Hours 195 244 735 Days 8 10 31 Weeks 1.2 1.5 4.4

9Planning Mode Applications The planning mode can be used when the policies or strategies being considered are expected to result in regional-scale changes in overall levels of travel demand or changes in regional travelers’ destination, mode, or time-of-day choices, but are not expected to be significantly affected by local-scale traffic dynamics. The planning mode can be applied to the following primary policy and strategy analyses: • Pricing. Pricing strategies are the costs imposed on travelers using certain roads, traversing certain screenlines, or traveling to certain areas (tolling, cordon pricing, or area pricing). These costs may either be fixed or vary by time of day or by response to congestion. Additionally, these costs may vary by user to reflect discounts or subsidies provided to some users. Pricing strate- gies are most effectively addressed in the context of a regional-scale model given the potential responses to pricing strategies. Responses may include changes in the overall level of activity and trip generation, changes in the destinations for these activities, changes in the travel modes used to access these destinations, and changes in the specific routes on the roadway or transit networks given the selected mode. • Capacity. Capacity strategies involve adding, modifying, or deleting capacity on the roadway system. This may include the addition of new roads or lanes to the travel model networks, or it may involve adjusting existing capacity, such as the implementation of reversible lanes, auxiliary lanes, or turn lanes at intersections. The impacts of local capacity enhancements may be better captured using traffic microsimulation tools, but significant increases in capacity (such as the addition of new roads or lanes) are better addressed using regional-scale models because the enhancements have potentially broad impacts on regional network levels of ser- vice, which could influence generation, distribution, mode choice, and route assignment. • Travel demand management. Travel demand management strategies typically aim to change travel behavior to reduce congestion and improve mobility. For example, these policies may seek to increase the number of people who work at home and their frequency of doing so; to induce workers to adjust their schedules to travel during off-peak, less-congested conditions; or to increase the number of people who carpool to work. Such policies are most appropriately addressed at a regional scale because of their expected impact on performance on regionally significant or congested facilities. However, a detailed model of traffic dynamics is not neces- sarily required to capture the impact of these policies. Network impedances (VDF-based) Demand (activities and trips) TRANSIMSSTUDIO Iteration/Convergence FileManager DaySim Exogenous Trips TRANSIMS Router MOVES MOEs / Indicators TRANSIMS Router Microsimulator Figure ES.3. Planning mode system configuration.

10 Operations Mode The operations mode can be used when the analysis requires an assessment of the impacts of a policy or strategy on local traffic dynamics and when these improvements are not expected to result in changes in overall levels of travel demand or in destination, mode, or time-of-day choices. Operations Mode Configuration The distinguishing features of the operations mode are that it incorporates a full regional traffic microsimulation but does not include an iterative feedback loop in which microsimulator- based network simulation impedance measures are fed back to DaySim. The elimination of this feedback loop reflects the fact that some operational improvements may greatly improve local traffic dynamics but have only marginal effects on the other travel dimensions; it also acknowl- edges that regional microsimulation is computationally intensive and results in extremely long runtimes. Figure ES.4 illustrates the configuration of the model system components in the operations mode. Operations Mode Application The operations mode is most appropriate when the policies or strategies under consideration are not expected to result in significant changes in overall levels of travel demand or in destination, mode, or time-of-day choices. The operations mode can be applied to the following primary policy and strategy analyses: • Capacity. Some capacity improvements or enhancements to existing capacity may be evaluated using more local, operationally focused tools. These may include changes such as turn lanes at intersections, other geometric changes such as lane connectivity or lane widths, and the pres- ence of shoulders. • Operations. The operations model may support the analysis of bottleneck improvements, such as the addition of new signals or signs, adjustment of signal timing and phasing, or implementation of ramp meters. These are often most appropriately tested at a local scale using network assignment tools while holding the other choice dimensions fixed. However, more extensive bottleneck or other operational improvements—such as those applied across an extensive, coordinated signal system within a corridor or at a regional scale— may be appropriately tested using the planning + operations application mode described as follows. Demand (activities and trips) TRANSIMSSTUDIO Iteration/Convergence FileManager DaySim Exogenous Trips TRANSIMS Router Microsimulator MOVES MOEs / Indicators Figure ES.4. Operations mode system configuration.

11 Planning + Operations Mode The planning + operations mode represents the fully integrated DaySim and TRANSIMS model system. In this application mode, the TRANSIMS Router and Microsimulator are used to perform a regional traffic microsimulation as part of every model system global iteration; microsimulation-based network impedance measures are fed back and used as input to DaySim. The advantage of this application mode is that it provides the full range of sensitivities to changes in both regional demand and local and regional traffic dynamics. However, these extensive sen- sitivities come with the significant disadvantage of extremely long model system runtimes. Figure ES.5 illustrates the configuration of the model system components in the planning + operations mode. Planning + Operations Mode Applications The planning + operations mode can be used when the proposed policies or strategies are expected to result in regional-scale changes in the overall level of travel demand, or changes in regional travelers’ destination, mode, or time-of-day choices and are expected to be significantly affected by local-scale traffic dynamics. The planning + operations mode can be applied to the following primary policy and strategy analyses: • Pricing. Pricing strategies are the costs imposed on travelers using certain roads, traversing certain screenlines, or traveling to certain areas (tolling, cordon pricing, or area pricing). The costs may be fixed or vary by time of day, or they may respond to congestion. Additionally, the costs may vary by user to reflect discounts or subsidies provided to some users. Pricing strate- gies are most effectively addressed in the context of a regional-scale model given the potential responses to pricing strategies. Responses may include changes in the overall level of activity and trip generation, changes in the destinations for these activities, and changes in the travel modes used to access these destinations. The planning + operations mode is necessary when the pricing policies are also expected to have regionally significant impacts on traffic dynamics. • Capacity. Capacity strategies involve adding, modifying, or deleting capacity on the roadway system. This may include the addition of new roads or lanes to the travel model networks, or it may involve adjusting existing capacity, such as the implementation of reversible lanes, auxiliary lanes, or turn lanes at intersections. Significant increases in capacity (such as the addition of new roads or lanes) are most effectively addressed using regional-scale models. Figure ES.5. Planning + operations mode system configuration. Network impedances (Simulation based) Demand (activities and trips) TRANSIMSSTUDIO Iteration/Convergence FileManager DaySim Exogenous Trips TRANSIMS Router Microsimulator MOVES MOEs / Indicators

12 These models have potentially broad impacts on regional network levels of service, which could influence generation, distribution, mode choice, and route assignment. The use of regional traffic microsimulation within the integrated model provides a more robust platform for estimating these impacts. • Travel demand management. Travel demand management strategies typically aim to change travel behavior to reduce congestion and improve mobility. For example, these strategies may seek to increase the frequency and numbers of people who work at home; to induce workers to adjust their schedules to travel during off-peak, less-congested conditions; or increase the number of people who carpool to work. Such policies are most appropriately addressed at a regional scale because of their expected impact on performance on regionally significant or congested facilities. Although a detailed regional-scale model of traffic dynamics is not neces- sarily required, the microsimulation may provide a better tool for assessing the impact of these policies. • Operations. The planning + operations model can support the analysis of bottleneck improve- ments, such as the addition of new signals or signs, adjustment of signal timing and phasing, or implementation of ramp meters. These are often tested at a local scale using network assign- ment tools while the other choice dimensions are held fixed. However, more extensive bottle- neck or other operational improvements—such as those applied across an extensive, coordinated signal system within a corridor or at a regional scale—may be appropriately tested using the fully integrated planning + operations application mode. • Greenhouse gas. Strategies to reduce greenhouse gases (GHGs) may include both land-use and transportation improvements. In this case, the project team focused on transportation strate- gies. Using the microsimulator integrated with DaySim in the planning + operations mode provides the greatest sensitivity to GHG reduction strategies; these include pricing strategies to reduce VMT or increased fuel-efficiency standards. In addition, integration produces the most detailed estimates of transportation measures used as inputs to the MOVES and GHG estimation tools. Before using the model system in an application, the DaySim and TRANSIMS model compo- nents had to be implemented and linked. In the Jacksonville region, the model system was then calibrated and validated. These efforts, as well as the initial sensitivity testing of the model system in Burlington, are the focus of this report. Conclusions Model Implementation Demand Model Data Development Developing the parcel-level inputs to the activity-based model (ABM) was relatively straight- forward. Cleaning the employment data by the North Florida Transportation Planning Organization (NFTPO) and the Florida Department of Transportation (FDOT) significantly reduced the amount of time required to implement the model, although relatively crude updates to the employment data in one of the counties were still necessary. The parcel file required some additional cleaning to establish reasonable totals of housing units and to address inconsistencies in the parcel geography. School enrollment, transit stops, intersection types, and parking data were all relatively easy to assemble from existing data sources. In addition, developing the syn- thetic population was relatively straightforward given the availability of the data and tools; how- ever, the overall effort still required approximately 6 months. Accommodating auxiliary demand within the integrated DaySim-TRANSIMS model system was achieved using readily available static methods from the region’s trip-based model; how- ever, revisions to these auxiliary demand components are necessary for a more spatially and temporally consistent integrated demand-supply model system. A drawback of the current

13 implementation is that the auxiliary demand is fixed for each forecast year. That is, although this demand varies by forecast year, it is not affected by changes in network impedances. Ideally, the auxiliary models will be revised to provide sensitivity to changes in network performance. Network Model Data Development Developing detailed and usable networks for microsimulation requires significant effort. The TRANSIMS software comes with a wide array of tools to perform many network development tasks, and spatially detailed network data are widely available. However, users should expect to spend hundreds of hours debugging simulation networks, by correcting topological errors, resolv- ing attribute discontinuities, and coding intersection controls. The time-consuming effort involves iteratively evaluating, adjusting, and testing the networks by running simulations. In addition, users face numerous challenges when attempting to develop future-year or alternative network scenarios, a topic discussed in subsequent sections. Model Integration Configuring DaySim to generate temporally, spatially, and behaviorally detailed travel demand information for use in TRANSIMS was straightforward. Configuring TRANSIMS to generate the skims for input to DaySim was also straightforward. More sophisticated methods of providing TRANSIMS-based impedances to DaySim could potentially be implemented. These could include implementing efficient multistage sampling of destinations (and corresponding impedances) at strategic points in the DaySim looping process, or tightly integrating DaySim and TRANSIMS so that DaySim can call TRANSIMS to extract the required measures quickly. However, the project team decided that the runtime implications and resources required for development were pro- hibitive, and concluded that the current methods provide sufficient spatial and temporal detail. The network convergence equilibration effort revealed that the most effective convergence strate- gies were often the least acceptable to the larger DTA community, but they were necessary to ensure sufficiently converged assignments within reasonable runtimes. Schedule consistency was identified as another measure of the soundness of a model solution. Extensive testing of the model system was necessary to determine the number of network assignment and model system itera- tions required to ensure that differences between alternative scenario model results were attribut- able to these policies and investments and were not obscured by noise in the model system. Model Enhancements The enhancements made to the model system were necessary to improve the model system’s sensitivity and to fulfill the goals of the SHRP 2 C10A project. Updates to the DaySim model system were relatively straightforward, although the updates were not fully completed until a new DaySim software architecture was implemented, which took significantly longer than expected. The updates to the TRANSIMS model components were much more extensive and involved much more time to implement; many of these enhancements were under development during the C10A project. While these enhancements were necessary to fulfill project goals, they undoubtedly also resulted in schedule delays. Model Application The challenges in interacting with the model are primarily associated with debugging the model system. As already mentioned, the network simulation model is very sensitive to small-scale network coding and parameter assumptions, and the network simulation is subject to frequent failures as input assumptions are refined. Users must be able to understand and mine which data generated by the model system can illuminate the source of simulation problems and also be able

14 to make informed decisions about how to modify model inputs to achieve the proper model sensitivity. Model users must also have a basic understanding of Python programming language to understand the overall model system flow, as well as robust data manipulation, statistical analysis, and geographic information system (GIS) skills. Model users are not required to know C# or C++, the development platforms used for DaySim and TRANSIMS, respectively. The types of analysis that can be performed with the new model system are fundamentally different and more expansive than can be performed with a traditional model system, and the application and interpretation of model outputs must be thoughtfully considered. The fully integrated model system is most valuable when the proposed policies or strategies are expected to result in regional-scale changes in the overall level of travel demand or changes in regional travelers’ destination, mode, or time-of-day choices, and are expected to be significantly affected by local-scale traffic dynamics. The model system software can be flexibly deployed on hardware running either Windows or Linux, and the implementation can be scaled or configured to reflect available hardware resources. To avoid long runtimes, the model can be used in different application modes, as described earlier in this chapter. Although many DaySim and TRANSIMS tools exist to assist in data preparation and coding, the model system is highly sensitive to alternative configurations of the model system and to small-scale coding issues; anywhere from an hour to many weeks may be needed to generate plausible alternative scenarios. Model System Calibration and Validation Transferring the DaySim activity-based demand component from Sacramento (where DaySim was initially implemented) to Jacksonville radically reduced the amount of time needed to implement the activity-based demand model component of the model system. Additional calibration and validation of some of the subcomponents of the model—such as the daily activity pattern component of DaySim or the refinement of TRANSIMS networks—was nec- essary to improve model performance. However, a number of the models required little, if any, recalibration. The project team used the National Household Travel Survey (NHTS) as the primary observed data source for developing demand model calibration targets, although the survey’s limited weekday sample size in the Jacksonville region and other data completeness issues created some challenges. Ultimately, more time was spent refining and validating the roadway networks than refining the calibration of the DaySim demand model components. Network microsimulation models are significantly more sensitive to network coding assumptions, so identifying and resolving those issues simply require more time. Model Sensitivity Testing Travel demand forecasting model systems are only able to test the effects of policies and assump- tions that have been explicitly included in the design and implementation of the model system; they are not intrinsically sensitive to the increasingly broad range of transportation policies and improvements of interest to decision makers. While most regional models are sensitive to large- scale assumptions about land use and demographics, few are sensitive to more detailed assump- tions about pricing policies, or to traffic or travel demand management strategies. Even when models have the capability to address these types of policies, they are typically not sufficiently sensitive to the dynamic interplay between travel behavior and network conditions by time of day to do so, nor can they reasonably represent the effects of road pricing, travel demand manage- ment, and other policies. Sensitivity testing of model systems involves the evaluation of the effects of changes in model inputs on model outputs. The Burlington implementation of the C10A model system was subjected to a set of sensitivity tests designed to illustrate the unique capabilities of the model system, including pricing, travel demand management, and operations.

15 Pricing Two types of pricing tests were evaluated as part of this effort. In the first, a number of scenarios were defined in which freeway tolls varied by time of day. In the second, scenarios were defined in which auto operating costs were modified from the baseline condition. For the first type of sensitivity tests, three scenarios were evaluated and compared with the baseline. In the baseline alternative, no costs were assessed at any time; in the three scenarios, different fixed, per-mile charges that varied by time of day were evaluated. The expected responses to these policies—that travel would decline during tolled periods and on tolled facilities and that the changes would vary by activity purpose—were all observed in the model system outputs. Interestingly, all three pricing scenarios resulted in pronounced increases in travel demand during the evenings, suggesting that travelers will reschedule activities to occur when tolls are absent and when they have fewer sched- uling constraints such as are present during midday. The team also observed in these tests that tolls have different effects on different trip purposes. For example, work-related travel was rela- tively unaffected, but social- and recreation-related travel shifted noticeably out of the peaks and into the evening. Finally, the network-based total delay was higher than the base in all scenarios, as tolling induces travelers to shift onto more capacity-constrained surface facilities. For the second set of pricing sensitivity tests, three auto operating cost scenarios were evaluated and compared with the baseline. The baseline alternative assumed a cost of $0.12/mi, while the alternatives assumed charges as low as $0.06/mi and as high as $0.60/mi. These tests confirmed that when auto operating costs decline, the share of households choosing to maintain zero vehicles also declines; and as the costs increase, the share of zero-vehicle households also increases. How- ever, these changes were relatively modest. The results also showed small changes in regional tour frequency by purpose, although these shifts did not result in significant changes in network per- formance or congestion. Travel Demand Management Travel demand management (TDM) approaches incorporate a wide range of strategies aimed at changing travel behavior to reduce congestion and improve mobility. The sensitivity testing for Project C10A focused on assessing the impacts of a flexible work schedule in which all workers worked fewer days but longer hours on those days. The overall time spent on work activities was held fixed. The model results were consistent with expectations based on the structure and link- ages of the DaySim and TRANSIMS model. In general, overall levels of activity generation were lower, although the declines in work-related travel were offset by increases in travel for discre- tionary purposes. The model produced shifts in the distribution of travel by time of day because of the lengthened workday; as expected, changes in the destination and mode choices were rela- tively small. This test did reveal noticeable changes in network performance, with reduced con- gestion across all facility types throughout most of the day. A slight increase in congestion in the evening reflects both the later return times from work and increased participation in discretion- ary activities in the evening. Operations The sensitivity testing focused on a scenario in which signals were coordinated using TRANSIMS tools along three primary regional corridors, with the goal of reducing bottlenecks and improv- ing the overall traffic flow. The DaySim-TRANSIMS model system provides sensitivity to these improvements. (Traditional travel demand forecast models cannot typically represent such improvements because of their linkage with traditional static network assignment methods; those methods lack detailed network operation attributes and have coarse temporal resolution.) The initial model results showed some reductions in delay by facility type, particularly during peak periods. However, closer inspection of the speed profiles along the three targeted corridors

16 showed more mixed results, with the signal progression producing better speeds in some corri- dor directions and worse speeds in other corridor directions. As others have noted, the sensitivity of DTA and traffic microsimulation models to these detailed inputs suggests distinct challenges when attempting to incorporate these assumptions into a forecasting mode, especially at a regional scale. Of all the scenarios evaluated as part of this sensitivity testing, the signal progres- sion scenario required the greatest amount of time and resulted in the least interpretable results. Disaggregate Framework Because both the demand and the supply components of the model system are fully disaggregate, users can trace the impacts of policies and investments on individual travelers from long-term choices (such as usual work locations) all the way down to the specific paths taken by each indi- vidual traveler on a second-by-second basis. Although disaggregate model results are not reported, this framework provides tremendous flexibility for aggregating model results for specific travel markets or communities of concerns; and it is useful for debugging, calibrating, and refining model sensitivity. Also, note that random simulation variation did not compromise the ability to use the model system, provided that sufficient convergence was achieved both within the net- work assignment and for the model system overall. Overall, the new model system is more sensitive to a wider range of policies than a traditional travel demand model system. This sensitivity is further enhanced by the detailed representation of temporal dimension, as well as the increased behavioral and spatial detail. In addition, the model system produces a wider range of statistics of interest to decision makers. Extracting, man- aging, and interpreting the results was not difficult; however, the level of effort required to effec- tively test different types of improvements varied widely, from as little as 1 hour to more than a week. Using the model to evaluate the pricing and TDM scenarios was relatively easy, requiring straightforward adjustments to network coding or to model coefficients. Using the model to eval- uate the operational scenarios required significantly more effort because of the sensitivity of the network simulation to different signal coordination and timing assumptions. This level of effort would undoubtedly increase if more extensive changes to operational assumptions were required. In addition, even with the additional effort, the results produced by the model system did not seem as intuitive as the results of the other scenario tests.

Next: Chapter 1 Model Implementation »
Dynamic, Integrated Model System: Jacksonville-Area Application Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s second Strategic Highway Research Program (SHRP 2) Report S2-C10A-RW-1: Dynamic, Integrated Model System: Jacksonville-Area Application explores development of a dynamic integrated travel demand model with advanced policy analysis capabilities.

The report describes the implementation of the model system in Burlington, Vermont, and in Jacksonville, Florida; the calibration and validation of the model system; and the application of the model system to a set of initial sensitivity tests.

The same project that developed this report also produced a report titled Transferability of Activity-Based Model Parameters that explores development of regional activity-based modeling systems for the Tampa Bay and Jacksonville regions in Florida.

Capacity Project C10A developed a start-up guide for the application of the DaySim activity-based demand model and a TRANSIMS network for Burlington, Vermont, to test linking the demand and network models before transferring the model structure to the larger Jacksonville, Florida, area. The two model applications used in these locations are currently available.

Software Disclaimer: This software is offered as is, without warranty or promise of support of any kind either expressed or implied. Under no circumstance will the National Academy of Sciences or the Transportation Research Board (collectively "TRB") be liable for any loss or damage caused by the installation or operation of this product. TRB makes no representation or warranty of any kind, expressed or implied, in fact or in law, including without limitation, the warranty of merchantability or the warranty of fitness for a particular purpose, and shall not in any case be liable for any consequential or special damages.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!