Trip Generation Models PDF

# Session 5: Trip Generation ## Session Outline - Terminology - Key Concepts - Inputs and Outputs - Error Checking, Calibration, and Validation ## Objectives - Define productions and attractions by trip purpose. - Identify typical inputs into a trip production model. - Identify outputs from a trip...

# Session 5: Trip Generation ## Session Outline - Terminology - Key Concepts - Inputs and Outputs - Error Checking, Calibration, and Validation ## Objectives - Define productions and attractions by trip purpose. - Identify typical inputs into a trip production model. - Identify outputs from a trip production model. - Identify typical inputs to a trip attraction model. - Identify typical outputs from a trip attraction model. - Explain the logic for balancing productions and attractions. - Explain the logic for balancing productions and attractions by TAZ for NHB trips. ## Session Outline This session covers trip generation, the first step in the traditional TDF process. The session includes a discussion of trip purposes, a discussion of how the number of trip ends are calculated for each TAZ using zonal household and employment data. The session also includes a discussion of trip rate information developed from the travel surveys. The inputs and outputs to the trip generation step, the treatment of special generators, the balancing of attractions to productions, and error checking are also discussed. ## Terminology - Trip Generation - Person Trip - Vehicle Trip - Trip End - Trip Production - Trip Attraction - Home-based (HB) Trip - Non-home based (NHB) Trip - Special Generator ## Key Concepts ### Trip Generation Model - Estimates person trip ends for each TAZ - Equations predict trips based on attributes - Dependent variables are trip rates calculated from household and workplace surveys, and the TAZ household and employment attributes developed using demographic and employment estimation methodologies. ## Trip Generation Model Trip generation is the first step in the traditional four-step TDF process. The trip generation model estimates the number of person trip ends for each TAZ. These trips are calculated on a household, person, or employee basis and then aggregated to TAZ level. Therefore, trip generation uses disaggregate data and a disaggregate model to perform aggregate estimation or forecasting. The trip generation model consists of equations that are established to predict the number of trips generated based on TAZ household and employment data and regional trip rates developed from household and workplace travel surveys. ## Trip Purposes The image shows a table and diagram. The diagram describes the types of trips that occur between four different zones, one being a home, one being a work location, and two other locations. All trips end at home and all trips begin at home. | TAZ | P | A | P | A | P | A | |---|---|---|---|---|---|---| | 1 | 1 | 1 | 1 | 1 | 3 | 3 | P =Production A = Attraction HBW = Home-based work NHB = Non-home based HBO = Home-based other Practice has shown that better travel forecasting models are obtained if trips by different purposes are identified and modeled separately. The most common trip purposes are: - HBW - HBO - NHB In TDF, trip productions and attractions are used to represent the ends of a trip. A production is the home end of a HB trip and the beginning of a NHB trip. An attraction is the non-home end of a HB trip and the end of a NHB trip. ## Trip Generation Process The image shows a flowchart. The image describes how to conduct trip generation using zonal socioeconomic data and zonal employment data. ### Productions - Zonal Socioeconomic Data - Production Model - Household Travel Survey Data - Production Trip Ends by Purpose by TAZ ### Attractions - Zonal Employment Data - Attraction Model - Workplace Travel Survey Data - Attraction Trip Ends by Purpose by TAZ There is also a step in the flowchart called **Balancing Productions and Attractions by Purpose**. This step brings production and attraction trip ends together. It is followed by **Production and Attraction Table by Purpose**. ### Information Needs for Trip Productions Socioeconomic disaggregate data are required to develop trip production models. Typical socioeconomic data elements include: income, automobile availability, household size, and number of workers. Socioeconomic data from the Census and household activity data from travel surveys are used to develop trip production models. ### Information Needs for Trip Attractions It is difficult to obtain adequate information to develop trip attraction models. For trip attractions, the most important data element is employment. Employment is broken down into categories. The most common categories are basic employment, retail employment, and service employment. Trip attraction rate information is developed from workplace and household surveys. ### Special Generator Information Needs To obtain trip rates and trip estimates for special generators, specific information relating to each type of generator needs to be obtained. For example, in the case of a hospital, the number of beds is important. For a large shopping mall, the gross leaseable area is used. ### Trip Rates From Travel Surveys Travel surveys include household, workplace, external station, special generator, truck, and transit on-board surveys. The data obtained from travel surveys are used to develop trip production and trip attraction rates for each trip purpose. ## Trip Generation Models The image shows a flowchart of trip generation models. - Trip Rates - Zone Demographics The flowchart includes: - Home-based Work Productions & Attractions - Home-based Other Productions & Attractions - Non-Home Based Productions & Attractions ### Trip Generation Model Components - Trip production models - Trip attraction models - Estimation of special generator trip generation - Balancing productions and attractions Trips generally are stratified by purpose, which typically include: HBW, HBO, and NHB trips. HBO trips can be stratified further (HB school, HB shopping, etc.) depending on regional travel characteristics and data availability. In addition to this stratification, the TAZs can also be divided into various levels of residential densities, called area types. Area types include the CBD, CBD fringe, urban residential, suburban residential, and rural. HB trips in urban areas are about 70% of total person trip making. The trip generation rates typically includes trip production models, trip attraction models, estimation of special generator trips, and a procedure for balancing trip productions and attractions. The models are mathematical functions used to estimate trip ends. The outputs from the trip generation step are productions and attractions by trip purpose for each TAZ. ## Production Model ### Cross-classification Model A cross-classification model is one type of model used to determine trip productions. Regression models are also used for trip productions but are not illustrated here. This model is based on estimating the response (e.g., the number of trip productions per household for a given purpose) as a function of household attributes. The model assumes that trip production rates are stable over time for households. Trip rates are derived empirically from travel surveys with the same household characteristics. Cross classification is based on grouping the households in different strata; for example, a specific cell is based on household size and household income. The trip production rate for that specific cell then is the total number of trips in that cell divided by the number of households in that cell. This relationship can be stated as follows: Where: * h = Households with a particular combination of characteristics (for example a two-person family in income group 1) * t(h) = Trip rate for purpose p made by members of households of type h * TP(h) = Total number of trips by purpose p made by households in cell h * H(h) = Number of households in cell h * t'(h) = TP(h)/H(h) Following are some advantages of the cross-classification model: - Groupings are independent of the TAZ system of the study area. - No prior assumptions of the shape of the relationship are required and can easily accommodate non-linear relationships. - It can also be employed for mode split. - It is simple to use and to understand. Following are some disadvantages of the cross-classification model: - It does not permit extrapolation beyond its calibration strata. - There are no statistical goodness-of-fit measures for the model. - It requires large sample sizes (25 households per cell); otherwise cell values will vary in reliability. ## Attraction Model ### Cross-classification Models The table shows the number of attractions per different types of employment for Home based work (HBW) and non-home based (NHB) trips. **HBW** | Person Trips | Attractions per Employee | Attractions per Household | |---|---|---| | Basic | 1.60 | 0.082 | | Retail | 1.35 | | | Service | 1.39 | | **NHB** | Person Trips | Attractions per Employee | Attractions per Household | |---|---|---| | Basic | 0.99 | 0.361 | | Retail | 4.53 | | | Service | 1.63 | | HB trip attraction models usually are aggregate models for predicting trip ends that are associated with the non-home end of the trip. A lack of data is a major problem with trip attraction models. Household surveys provide excellent data for production models but much less information for attraction models. The same trip purposes used for the trip production models are used for the trip attraction models. Trip attraction models have received less attention than production models. Typically, cross-classification models are used, or regression models that relate trips to the various land use types to which they are attracted are used. In cross-classification, the most commonly used classification is employment type, which is divided into groups such as basic, retail, and service employment. Trips also are attracted to households. Household trip rates are developed from household travel surveys. The examples provided are trip rates developed from the 1998 Austin, Texas, travel surveys. ## Balancing Attractions to Productions The image shows a table of productions and attractions for HBW trips. The table demonstrates how attractions are adjusted to equal productions. Productions are set as 1,900; the initial attractions are set as 2,200. Attractions are multiplied by 0.8636, which brings them down to 1,900. The final step in trip generation modeling is balancing productions and attractions. The estimated number of trips produced by households should be equal to the number of trips attracted to activity centers. Each trip has a production and an attraction end. In practice, the estimation of productions and attractions will not be equal. Trip production totals normally are used as control totals, and attractions are scaled to productions because there is a greater degree of confidence in the production models than in the attraction models. **Rule of Thumb - Best practice suggests that production/attraction factors should be between 0.9 and 1.1.** ## Balancing Non-Home Based Trips The image shows a diagram of a trip between two zones and shows that, for a non-home based trip, production and attractions are balanced to equal each other. NHB trip production rates were developed for the "home" zone or travelers, but by definition, for NHB trips neither end is at home. Therefore, we discard the NHB productions generated in the "home" TAZ and substitute them with the NΗΒ attraction for that TAZ, assuming that people attracted to that TAZ will be produced out of that TAZ. This procedure means that NHB trips must balance by TAZ; for example, if TAZ 1 has 100 NHB attractions, it must also have 100 NHB productions. ## Error Checking, Calibration, and Validation - Compare trip percentages by purpose - Compare person trips per household by purpose - Check treatment of special generators Error checking of the trip generation step includes comparing the results to other models and to state or federal averages for consistency and reasonableness. Application of the trip generation model for a previous year for which survey data are available may provide a test for the model’s stability. Carefully scrutinize the following: - Compare trip percentages by purpose to percentages from other surveys, other urban areas, or typical values: | Typical Values | |---| | HBW | 20% | | HBO | 57% | | NHB | 23% | Source: NCHRP Report 365, Travel Estimation Techniques for Urban Planning, 1998 - Compare person trips per household to other surveys, other urban areas, or typical values. | NBW | 1.70-2.30 trips per household | |---|---| | HBO | 3.40-4.80 trips per household | | NHB | 1.90-3.00 trips per household | Source: Model Validation and Reasonableness Checking Manual, TMIP, 2001. - Ensure that facilities such as shopping malls, military installations, universities, and airports, which have high or low trip attraction rates per employee, are represented correctly in the trip generation model. ## References 1. Ortuzar J. de D., and L.G. Willumsen. Modeling Transport. 2nd ed. West Sussex, England: John Wiley & Sons, 1994. 2. Institute of Transportation Engineers. Trip Generation. 6th ed. Washington D.C.: 1997. 3. Harvey G, and E. Deakin. A Manual of Regional Transportation Modeling Practice for Air Quality Systems. National Association of Regional Councils, 1993. 4. Urban Travel in Texas: An Evaluation of Travel Surveys. Research Report 1099-3F. Texas Transportation Institute, The Texas A&M University System, College Station, Texas, January 1996. 5. Model Validation and Reasonableness Checking Manual, TMIP, June 2001. 6. Travel Estimation Techniques for Urban Planning, NCHRP Report 365, 1998.