PreDay

Pre-day model framework

The pre-day model follows the Day Activity Schedule approach (1, 2) and is formulated through a system of interconnected discrete choice models focusing on decisions related to daily activity and mobility. The overall model structure, overview of models of different levels, accessibility measures and the data used for development and estimation will be covered in this section (for a full review see 3).

Practical instructions on how to run PreDay can be found here.

Practical instructions on how to modify pre-day: add new modes, add new activities, make repeatable preday and configure operational costs please see this page.

The pre-day model is a disaggregate travel demand system designed to model the decisions at mid-term level. The long-term decisions are input to the system and thus are exogenous. Figure 1 and Figure 2 show the model components and process flow of the pre-day model, respectively. Synthetic population with known socio-demographic characteristics is an input to the system. Other inputs include network skims, land use characteristics, etc. Those inputs are discussed subsequently. There are three different hierarchies in the system: day pattern level, tour level and intermediate stop level. Each level consists of several models. Given the complicated and multi-dimensional nature of the day activity schedule problem, curse of dimensionality will occur if every choice dimension is put into one single choice model. Therefore, choice levels are isolated, structured, and put into different choice models. The overall system can be viewed as a hierarchical (or nested) series of choice models. For a detailed description of the models, including the technical reports, see SimMobility Mid-Term Module Parameters page: https://github.com/smart-fm/simmobility-prod/wiki/Mid-Term-Parameters Figure 2 highlights the hierarchical structure of the model. The solid arrows indicate that models from lower levels are conditioned on decisions made with models from higher levels. The dashed arrows represent the accessibility measures described subsequently.

Figure .1: Pre-day activity-based travel demand model: Components

Figure .2: Pre-day activity-based travel demand model: Process flow

Day pattern level

This level distinguishes the pre-day model (an activity-based model) from tour-based models (see for example 4 and 5) because it organizes tours and manages their sequence through the concept of day activity schedule. Day activity schedule is defined through the concepts of activity pattern, and activity schedule. Activity pattern defines the participation in activities as primary, and secondary. Primary activities are the anchors (e.g., home to work trip, and work to home trip represent a tour with work as primary activity) of tours, and secondary activities are intermediate stops within a particular tour (e.g., stopping for shopping at a work-to-home half-tour). Activity schedule adds detailed information about tours to the activity pattern such as sequence, timing, travel mode, destination of primary activity, and also the stops for secondary activities within tours.

In the pre-day activity-based travel demand model, the day pattern level (see Figure 1) includes two types of discrete choice models: day pattern model, and exact number of tours model for different primary activity purposes. The day pattern model predicts occurrence of tours for various purposes and availability of intermediate stops for various purposes. The purposes are defined by four activity types: work, education, shopping, and others. Tour purposes that are predicted to occur will be passed to a second model to determine the exact number of tours for that purpose. The predicted availability of intermediate stops has no immediate effect at day pattern level. However, the results will be provided to intermediate stop generation model to constrain the availability of each activity purpose. Day pattern level will generate a list of tours as well as intermediate stop availabilities for each individual in the synthetic population. Stop availability is set according to person’s type. For example, for non-full time students, education tours will not be available.

Tour level

A tour is defined by a set of trips with the origin of the first trip and the destination of the last trip being home. In other words, tours are home-based, except for tours predicted by the work-based sub-tour model, which are work-based (tours that start from work and end at work). In the pre-day activity-based travel demand model, the tour level includes multiple discrete choice models: usual/unusual work location; travel mode choice or travel mode/destination choice; work-based sub-tour generation; tour time of day. These models provide detailed information for each predicted tour. These details include destination, travel mode, time of day (arrival time and departure time). For every work tour, there is a specific model at this level to determine whether a sub-tour is going to be scheduled. For the purpose of modeling the tour-level decisions for work-based sub-tours, two additional sub models for work-based sub-tours, namely sub-tour mode/destination choice and sub-tour time-of-day choice, should be specified. Sequencing of tours is accomplished before modeling individual tours by assigning each of the tours predicted at day pattern level a priority number. The priority number is determined by the purpose of tour primary activity. It should also be noted that the time of day models for tours are based on the concept of cyclical and continuous indirect utility functions (6). In summary, this level provides activity and travel information for tours.

Intermediate stop level

A basic tour contains only 2 trips: the first one from home to tour primary activity and the second one from tour primary activity to home. A more realistic tour structure should consider the existence of intermediate stops during a tour. The intermediate stop level will generate these stops. Trips for secondary activities are represented as intermediate stops within a tour, and the available types of secondary activities have been predicted in the day pattern model. The intermediate stop level (see Figure 2) includes three types of discrete choice models: intermediate stop generation, mode/destination, and time of day. These models first generate intermediate stops for each tour and then predict the timing and destination of stops for secondary activities, as well as the travel mode. After applying the intermediate stop level models to the synthetic population, a daily activity schedule is generated for each individual in the population. The generated activity schedules provide the timing (arrival time and departure time) of each activity at a resolution of 30 minutes, the destination at zonal level and the travel mode for each trip/tour from a list of considered modes. The output is then fed to the within-day model to generate trip chains ready for simulation.

Accessibility measures

Disaggregate utility-based accessibility measures (7) originated from random utility theory are included within the pre-day activity-based modeling framework. These measures represent the expected maximum utility of a set of alternatives from a choice set of a discrete choice model and are consistent with random utility theory. In a hierarchical modeling system, accessibility measures are essential to capture the sensitivity of activity and travel decisions modeled in lower levels of the modeling hierarchy. In formal nested modeling hierarchies, such as the one for the pre-day model, the upward integrity comes from the composite measure of expected utility across the lower level alternatives, or the so-called “logsum”. The "logsum", the log of the denominator of this logit choice probability, gives the expected utility from a choice (out of a set of alternatives), and can be used to link different choices. The accessibility measures, or logsums, introduced in the pre-day model are shown in Figure 3 with dashed arrows. The pre-day model adopts a simple accessibility measure structure where disaggregate measures from tour mode or mode/destination choice models are fed directly to choice models in the day pattern level. Note that in the current implementation the Exact Number of Tours models logsums are not being passed to the day pattern level. The Activity-Based accessibility measure (ABA), a top measure of accessibility, obtained at the top of the hierarchy from the day pattern binary model, captures the relative attractiveness of various alternatives and can be used in project evaluation, as it expresses the consumer benefits.

Figure .3: Pre-day calculations process of accessibility

Data

This section describes the data used for the development and estimation of the components that make up the pre-day activity-based demand model.

Network skims

The level of service variables used in estimating the mode choice models come from network skims. The network skims are provided by the Land Transport Authority (LTA) in Singapore. Those skims contain zone-to-zone travel time, travel distance, and public transit fares matrices for 3 pre-defined time periods: morning peak from 7:30 am to 9:30 am, evening peak from 17:30 pm to 19:30 pm and off-peak hours for the rest of the day. In pre-day model, the time of day model is designed to capture the timing of trip/activity at a resolution of 30-minute interval. To generate a more realistic travel time for every 30-minute interval, GPS-enabled travel time data as well as traditional survey-based travel time data, and applies data fusion techniques were used for the whole day for cars. Similar techniques are adopted to generate travel time for every 30-minute interval for public transportation by using transit data collected with smart cards (for more information see 3).

Land use data

Land use data is important for the modeling of destination choice. This data provides information used to build attraction variables (8) included in destination choice models. Currently, the destination choices within the pre-day model are as precise as traffic analysis zones, or so-called MTZs in Singapore. A list of land use parameters is attached to each MTZ in Singapore.

Household travel survey data

For the estimation of each individual component in the pre-day model, Singapore Household Interview and Travel Survey 2008 or HITS2008. currently, the population that is fed into pre-day is a synthesized population from household interview travel survey 2012 data (HITS 2012).

References:

1. Bowman, J. L., 1998. The day activity schedule approach to travel demand analysis. Ph.D. thesis, Massachusetts Institute of Technology.

2. Bowman, J. L., Ben-Akiva, M., 2001. Activity-based disaggregate travel demand model system with activity schedules. Transportation Research Part A: Policy and Practice 35 (1), 1–28.

3. SIYU, L. (2015). ACTIVITY-BASED TRAVEL DEMAND MODEL: APPLICATION AND INNOVATION (Doctoral dissertation).

4. Algers, S., Daly, A., Kjellman, P., Widlert, S., 1995. Stockholm Model System (SIMS): Application. Volume 2: Modelling transport systems. In: Proceedings of the 7th World Conference on Transport Research, Sydney, Australia. pp. 345–361.

5. Cascetta, E., Biggiero, L., 1997. Integrated models for simulating the Italian passenger transport system. In: Transportation Systems 1997: A proceedings volume from the 8th IFAC/IFIP/IFORS Symposium, Chania, Greece. Vol. 1.

6. Ben-Akiva, M., Abou-Zeid, M., 2013. Methodological issues in modelling time-of-travel preferences. Transportmetrica A: Transport Science 9 (9), 846–859.

7. Ben-Akiva, M., Lerman, S. R., 1985. Discrete choice analysis: Theory and application to travel demand. Cambridge: MIT Press.

8. Daly, A., 1982. Estimating choice models containing attraction variables. Transportation Research Part B: Methodological 16 (1), 5–15.