Travel Demand Modeling - MIT OpenCourseWare

Travel Demand ModelingMoshe Ben-Akiva1.201 / 11.545 / ESD.210Transportation Systems Analysis: Demand & EconomicsFall 2008

Review Discrete Choice Framework– A decision maker n selects one and only one alternative i from a choice setCn {1, ,Jn}–Random Utility Model whereU in Vin (attributes of i, characteristics of n, β ) ε in Discrete Choice Models– Multinomial Logit– Nested Logit Correlated Alternatives Multidimensional ChoiceNext Travel Demand Modeling2

Outline Introduction Approaches– Trip– Tour– Activity Emerging Approaches3

Long Term Choices Urban Development Firm location and relocation decisions Firm investment in information technologyMobility and Lifestyle Decisions Labor force participation Workplace location Housing Automobile ownership Information technology ownership and access Activity program4

Activity and Travel Pattern Choices Activity sequence and duration Priorities for activities Tour formation Telecommunications options Access travel information––––Traffic conditionsRoute guidanceParking availabilityPublic transportation schedules Reschedule activities Revise travel plans5

Modeling FrameworkLand Use and EconomicDevelopmentLong TermHousehold & IndividualBehaviorLifestyle and Mobility DecisionsActivity and Travel SchedulingImplementation and ReschedulingTransportation SystemPerformanceShort Term6

The Fundamental Modeling Problem Adequately represent a decision process that has an inordinate numberof feasible outcomes in many dimensions Example - Activity ScheduleN um b er of activitiesS eq u enc eT im ingL oca tionM o deR outeT o tal N u m b er o fA c tivity S ch e d u le A lte rn ative s101 0 p er activity1 00 0 per activ ity5 pe r activity1 0 p er activity1010!1 001 0,0 0 0501 001 0 17 Simplify Achieve valid results7

Simplifying the Problem Discrete time intervals Individuals defined by socioeconomic variables Divide space into zones Categories of activities Depiction of travel patterns trips, tours, activity schedules8

Approaches to Modeling Travel Trip-based Integrated trip-based Tour-based Activity schedule9

Representing Activity/Travel HDHTimeSHH: HomeW: WorkHHHDTimeWSSDDTimeS: ShopD: Dinner out10

Trip-Based: The 4-Step ModelTrip PurposeHome-based work (HBW)Home-based shop (HBS)Home-based other (HBO)Non-home-based (NHB)Behavioral Steps1.2.3.4.Trip Generation (Frequency)Trip Distribution (Destination)Modal Split (Mode)Assignment (Route)11

The 4-Step Model: Trip Generation Trip Production Household Size, Household Structure, Income, CarOwnership, Residential Density, AccessibilityTrip Attractions Land-use and Employment by Category (e.g. Industrial,Commercial, Services), AccessibilityCross Classification, Regression, Growth Factor12

The 4-Step Model: Trip Distribution Trip matrixGenerations123Attractions j TJijj123:i:I 3:TI3 D1D2D3 T1jT2jT3j:Tij:TIj T1JT2JT3J:TiJ:TIJ Dj DJ O1O2O3:Oi:OI iTij Tj13

The 4-Step Model: Trip Distribution Gravity ModelTij α i Oi β j D j f (Cij ) , i 1.I and j 1.J Tij Oi ,i 1.I Dj ,j 1.Jj Tiji Where,- f (Cij ) Function of the generalized cost of travelfrom i to j and- α i and β j are balancing factorsSolve iteratively for Tij , α i and β j14

The 4-Step Model: Modal Split LogiteVautoP ( auto ) Vauto Vtransite e Nested Logite µI NMP(NM ) µI NMe e µI M15

The 4-Step Model: Assignment Route Choice– Deterministic: Shortest Path, Minimum Generalized Cost– Stochastic: Discrete Choice (e.g. Logit) Equilibrium– Supply Side– User Equilibrium vs. System Optimal16

Limitations of the Trip-Based Method Demand for trip making rather than for activities Person-trips as the unit of analysis Aggregation errors:– Spatial aggregation– Demographic aggregation– Temporal aggregation Sequential nature of the four-step process Behavior modeled in earlier steps unaffected by choices modeledin later steps (e.g. no induced travel) Limited types of policies that can be analyzed17

Complexity of Work Commute (Boston)Simple Commute23%36%(no other activities)SimpleSimplehome40%SimpleworkComplex Commute(includes non-work exworkbankAll AdultsSource:Females with Males withChildrenChildrenBen-Akiva and Bowman, 1998, “Activity Based Travel Demand Model Systems,” in Equilibrium and AdvancedTransportation Modeling, Kluwer Academic.18

Complex Responses to PoliciesExample: Peak-Period TollPre-Toll SchedulePotential Responses to ShopTimeTime Peak Period(a) ChangeMode & PatternShopTime(b) ChangeTime & PatternTime(c) Work at HomeFigure by MIT OpenCourseWare.Source:Bowman, 1998, “The Day Activity Schedule Approach to Travel Demand Analysis,” PhD Thesis, MIT19

Modeling Travel at the Level of the Individual Classic 4-step– Trip Frequency– Destination Choice– Mode Choice– Route Choice Beyond 4-step– Time of Day– Integrated Trips– Tours20

Integrated Trip-Based Framework(e.g., MTC, STEP)Auto ownership--HomeBasedbased worktripstripsHomeWorkHome Based Other tripsNon-Home Based trips21

Highlights of Integrated Trip-Based System Key features– Disaggregate choice models– Models are integrated, via conditionality and measures ofinclusive value, according to the decision framework Key weakness– Modeling of trips rather than explicit tours22

Tour-Based Framework (e.g. Stockholm)Work ToursOther ToursBusinessPersonalBusinessShoppingOther23

Highlights of Tour-Based System Key features– Explicitly chains trips in tours– Validated and widely applied Key weaknesses– Lacks an integrated schedule pattern– Doesn’t integrate well the time dimension Data requirements– Same as for trip-based models24

Basics of Activity-Based Travel Theory Travel demand is derived from demand for activities Tours are interdependent People face time and space constraints that limit their activityschedule choice Activity and travel scheduling decisions are made in the contextof a broader framework– Conditioned by outcomes of longer term processes– Interacts with the transportation system– Influenced by intra-household interactions– Occurs dynamically with influence from past and anticipatedfuture events25

Activity Schedule SystemActivity and TravelActivity PatternTours26

Activity Pattern Replaces trip and tour generation steps of trip and tour-based models Models number, purpose and sequence of tours– Tours are interdependentTable removed due to copyright restrictions.Source:Bowman, 1998, “The Day Activity Schedule Approach to Travel Demand Analysis,” PhD Thesis, MIT27

Example of Activity PatternsPortland, ORTable removed due to copyright restrictions.Source:Bowman, 1998, “The Day Activity Schedule Approachto Travel Demand Analysis,” PhD Thesis, MIT28

Tours Primary Tour– Primary and secondary destinations– Timing– Modes Secondary Tours– Primary and secondary destinations– Timing– Modes29

Model StructureActivity Patternprimary activity/tour type,#/purpose secondary toursPrimary Tourstiming, destinationand modeSecondary Tourstiming, destinationand mode30

Highlights of Activity Schedule System Key feature– Integrated schedule Key weaknesses– Larger choice set Unrealistic behaviorally Computationally burdensome– Incomplete representation Coarse representation of schedule Coupling constraints31

Portland Activity-Based Model[570 Pattern Alternatives]Day Activity PatternHome Based Tours- Time of day- Primary destination- Primary modeWork-Based SubtourLocation of Intermediate Stops32

Preliminary Application Results 0.50/mile Peak Period Toll Shift in patternsType of Pattern byprimary activity% 8%3.3%All patternsSource:% before100.0%Bowman, 1998, “The Day Activity Schedule Approach to Travel Demand Analysis,” PhD Thesis, MIT33

Preliminary Application Results 0.50/mile Peak Period Toll Shift in work patternsType of Work PatternAt home0 sec tours1 sec toursSimple work tour0 sec tours1 sec toursComplex work tour0 sec tours1 sec toursTotal work patternsSource:% before% .3%100.0%-2.3%-4.7%-2.0%Bowman, 1998, “The Day Activity Schedule Approach to Travel Demand Analysis,” PhD Thesis, MIT34

Preliminary Application Results 0.50/mile Peak Period Toll Shift in work tour mode and chainingType of work tourDrive alone simpleDrive alone chainedOther simpleOther chainedTotal work .6%47.4%10.6%54.9%100.0%Bowman, 1998, “The Day Activity Schedule Approach to Travel Demand Analysis,” PhD Thesis, MIT35

Preliminary Application Results 0.50/mile Peak Period Toll Tour purpose and time-of-day effectsA.M. PeakP.M. PeakMiddayOutside PeakTotalSource:Percent change in total number ofhome-based 1.00%Bowman, 1998, “The Day Activity Schedule Approach to Travel Demand Analysis,” PhD Thesis, MIT36

Trends in Transportation Demand Modeling DATA:Massive OD Surveys Small-Scale Detailed Surveys MODELING METHODS:Aggregate Models Disaggregate ModelsStatic DynamicCanned Statistical Procedures Flexible Estimation of Models APPLICATION/FORECASTING:Mainframe User-friendly GIS, powerful PC SystemsAggregate Forecasting Disaggregate Forecasting(microsimulation) BEHAVIORAL REPRESENTATION:Homogeneous Heterogeneous (including demographics,attitudes and perceptions)Trips Activity Schedules37

Emerging Travel Modeling Approaches Activity and Trip-Chaining Models––––Activity time allocationLife cycle, household structure and roleTemporal variation of feasible activities over the dayDistribution of travel levels of service during the day Increased Travel and Information Choices– “No travel” options (tele-commuting, tele-shopping, etc.)– Information causes changes in departure time, mode and routechoice– Choice set formation38

MIT OpenCourseWarehttp://ocw.mit.edu1.201J / 11.545J / ESD.210J Transportation Systems Analysis: Demand and EconomicsFall 2008For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.

Basics of Activity-Based Travel Theory Travel demand is derived from demand for activities Tours are interdependent People face time and space constraints that limit their activity schedule choice Activity and travel scheduling decisions are made in the context of a broader framework