Travel Modelling Group Technical Advisory Committee September 11, 2013

Today’s Agenda Software Improvements 2001 / 2012 Base Network Progress New TMG Toolbox Tools Network Packages Network Comparison 2001 / 2012 Base Network Progress Methodology Coding of Transit-Only Lanes Presentation on GTAModel Version 4.0+ Design Features Roadmap Data Requirements

Software Improvements - Tools Analysis: Transit line boarding counts for comparison against TTS Zone adjacency matrix construction using zonal geometry Network Editing Tool for deleting transit stops occurring on highways Tool for matching network rotation Network optimization tool – converts cosmetic nodes to link shape vertices on a subset of links.

Software Improvements - NWP Network Package (NWP) File Format: Compressed packaging of a single scenario (modes, nodes, links, turns, lines, link shapes) for easy sharing. The Input / Ouput folder in the TMG Toolbox has tools for importing/exporting network packages. Compression ratio ~ 90% e.g., a network which is 10MB as batch files will only be 1MB. Also exports / imports extra attributes (Emme 4 only)

Software Improvements – Network Comparison Based on Network Correspondence File for a primary scenario and a secondary scenario All nodes in both scenarios get assigned a twin from the other scenario All links in both scenarios get assigned a list of twins from the other scenario to account for link splitting Twinning is based on node proximity, not node or link ID! Currently two tools use the correspondence file: “Flag Twinned Links and Nodes” computes extra attribute indices based on comparison “Copy New Elements” copies untwinned elements from the secondary scenario into the primary scenario

2011 / 2012 Network Progress Base 2011 Network is complete; still waiting on feedback GTFS stop locations are being used to update the 2011 base network to a 2012 base network with transit. New links have been identified by GTFS stop locations, and were created in 2012. The 2012 base network will contain a node for every GTFS stop grouping. Transit itineraries will be generated from GTFS data; with some data cleaning. The 2012 base network will support full-day modelling

Base Network Progress – Transit Only Lanes Current Example: Spadina, St. Clair The issue: This works well for network representation, but can become an issue when using GTFS stop locations to generate transit schedules. c u cs sl

Transit-Only Lanes: Proposed Solution Flag links with transit-exclusive ROW with new mode ‘X’ (Emme 4 permits 52 modes, so ‘X’ is a different mode from ‘x’). Manually / programmatically calculate transit vehicle equivalents A similar procedure can be used to implement transit speed updating This frees up the ‘L’ and ‘Q’ modes. For example, ‘L’ could be reserved for standard-gauage LRVs (e.g., Eglinton Crosstown line)

GTAModel V4.0 Eric J. Miller, Ph.D. James Vaughan & Peter Kucirek Travel Modelling Group, University of Toronto Technical Advisory Committee Presentation September 11, 2013 Toronto,

Key Features Microsimulation 24-hour weekday TTS2012 based “Continuous” generation of activity/trip start times 4 (maybe 5?) time periods for network assignments AM Peak Mid-day PM Peak Evening/night (maybe split?) TTS2012 based Activity/tour-based Same inputs as a 4-step model: Population & employment by zone Emme road and transit assignments to generate travel times/costs

Microsimulation (1) Zone population is synthesized into individual persons with specific age, employment status, occupation, school status and driver’s license. Persons assigned to households with a given number of cars. Zone employment is synthesized into individual jobs by occupation type.

Microsimulation (2) In V4.0, observed TTS distributions will be used for synthesis. Basically what is done in GTAModel 2.0, 2.5, GGH Model. Distributions can be updated on a scenario version. Future versions can implemented more sophisticated synthesis procedures (or, perhaps someday be driven by the demographic component of a land use model).

Microsimulation (3) Advantages: Processing a list of persons is faster than processing many large matrices. As soon as one starts to disaggregate trip-makers by occupation, age, etc. list-based calculations are far more efficient. E.g., matrix-based HBW work mode split calculations for approximately 2000 zones, 4 occupation groups, FT/PT workers = 4x106x4x2 = 32x106 nested logit model calculations to model mode choice for approx. 2.5x106 workers!

Microsimulation (4) Advantages, cont’d: Eliminates aggregation problems within the mode choice model Provides detailed distributions of behaviour and impacts by type of trip-maker (any type of aggregation of trips, etc. is possible). Simplifies model calculations.

24-Hour Modelling GTHA needs 24-hour modelling capability PM peak is now the dominant peak Energy/emissions calculations require 24-hour analysis Economic evaluation really requires 24-hour analysis Off-peak transit markets need to be analyzed.

2012 TTS It was originally proposed to build a 2006 version of the model and then update it once 2012 TTS data were available. Given the eminent release of the 2012 data we believe it makes sense to go directly to the 2012 model. Also, the Emme 2012 24-hour transit network is closer to readiness than the 2006 24-hour network and is a higher priority for completion for a variety of applications.

Activity/Tour-Based Model It was originally proposed to build a “simple” tour-based model as a logical extension of current GTAModel and GGH Model practice (which effectively generate simple H-Work-H and H-School-H tours). In attempting to design this model, however, significant complications soon arose wrt modelling shopping and other-purpose trips in a “simplified” way (i.e., it quickly became “not simple”). It also became clear that we already had a solution to this problem in the TASHA model that has been under development and testing at UofT for the past decade. We therefore propose to re-estimate TASHA using 2012 TTS data as the GTAModel 4.0 model system.

TASHA: Advantages Advantages of using TASHA include: Fully operational code already exists within XTMF, thereby minimizing the amount of new code that will need to be generated. TASHA deals with all trip purposes in a relatively simple, straightforward way but generates tours of arbitrary complexity in a computationally very efficient manner. Considerable experience with TASHA already exists that can be applied to the new version. We feel strongly that trying to invent a “simple” tour-based model represents an inferior solution and use of resources than a direct implementation of TASHA.

TASHA: Features Agent-based microsimulation; models both persons and households. Activity-based – generates 24-hour weekday out-of-home activity patterns. Tour-based: tours are emergent out of the scheduling of out-of- home activities. Mode choice is tour-based. Arbitrarily complex tours can be generated and handled efficiently. Household-based: detailed auto availability & allocation models determine mode choice decisions. Interfaces with both Emme and MATSIM (and, indeed, any network model). Can be used as a replacement for the first 3 stages in a 4-step model (with standard 4-step inputs) or as the travel demand component in an integrated transport-land use model system). Model parameters can all be developed from TTS data (i.e., no special surveys are required). In addition to the Toronto implementation, TASHA has been applied for research purposes in Montreal, London UK and Changzhou China.

TASHA Class Structure World Households Episode Distributions Spatial Representation Persons Person Projects Project Agenda Individual Activity Episodes Schedule Household Joint Activity Zones Distance Matrix Travel Time Matrices Travel Individual & Persons exist within households. This allows TASHA to deal explicitly with: Vehicle allocation Ridesharing Joint activities/trips Serve-dependent activities/trips

Project Types Person-Level Projects Work School Shopping Other Project Project Project Project Household-Level Projects Joint Shopping Joint Other Serve Dependent Project Project Project The current project structure is quite crude: it reflects that available data used to build the model (an ordinary one-day household travel survey).

Activity Scheduling Project 1 episode 1.1 episode 1.2 …. Project 2 episode 2.1 episode 2.2 …. Project N episode N.1 episode N.2 …. TASHA is an activity scheduling model in which individual activity episodes are generated and then explicitly scheduled. Out-of-home activity patterns and their associated trip-chains (tours) are thus “built from scratch” rather than selected from a pre-specified set of feasible patterns. Thus, travel patterns dynamically adjust to changes in transportation level of service, activity system “supply”, changes in household and personal constraints and needs, etc. … Day 1 Day 2 Day 3 Day 4 Day 5 Day 6 Day 7

Joint Activities Person 1 Person 2 Joint Shopping Activity: Duration: 2 hrs Location: The Mall Search for feasible joint time slot Day n Day n …. …. …. ….

Serve Dependents Child’s Schedule Adult 1 Schedule Adult 2 Schedule At-Home At-Home Take child to/from daycare Daycare Work At-Home At-Home Shopping At-Home

Vehicle Allocation within TASHA TASHA assigns household vehicles to drivers based on overall household utility derived from the vehicle usage. Drivers not allocated a car must take their second-best mode of travel.

Household Ridesharing Options in TASHA Within-household ridesharing is explicitly handled within TASHA. Drivers will “offer” rides to household members if a net gain in household utility is obtained and feasibility criteria are met.

= “Gap” in Project Agenda TASHA generates the number of activity episodes from a set of “projects” that a person (or household) might engage in during a typical weekday. It also generates the desired start time and duration of each episode. It then builds each person’s daily schedule, adjusting start times and durations to ensure feasibility. Travel episodes are inserted as part of the scheduling process. Activity Episode Frequency, Start Time and Duration Generation (a) Draw activity frequency from marginal PDF (b) Draw activity start time from feasible region in joint PDF (c) Draw activity Duration from feasible region in joint PDF Joint PDF Joint PDF Activity Frequency Start Time PDF Start Time Duration Activity Frequency Feasible Start Times Feasible Durations Scheduling Activity Episodes into a Daily Schedule Work Project Work School Project Other Project Other Shopping Project Shop 1 Shop 2 : : Person Schedule At-home At – Home Work Shop 1 Other Other At-home Shop 2 At-home = “Gap” in Project Agenda = Activity Episode = Travel Episode

Tour-Based Mode Choice Chain c: 1. Home-Work 2. Work-Lunch 3. Lunch-Meeting 4. Meeting-Work 5. Work-Home Tour-Based Mode Choice mN = mode chosen for trip N Drive Option for Chain c Non-drive option for Chain c m1 = drive m5 m4 m3 Sub-Chain s: 2. Work-Lunch 3. Lunch-Meeting 4. Meeting-Work m2 m1 TASHA’s tour-based mode choice model: Handles arbitrarily complex tours and sub-tours. without needing to pre-specify the tours Dynamically determine feasible combinations of modes available to use on tours. Modes can be added without changing the model structure. Cars automatically are used on all trips of a drive tour. Drive for Sub-chain s Non-drive for Sub-chain s m2 = drive m3 = drive m4 = drive m4 m3 m2 m5 = drive 28

Treatment of Time Models all out-of-home activities and trips for a 24-hour typical weekday 5 minute time increments are used for start times and durations/travel times Provides great temporal detail but is computationally very efficient (integer storage & calculations) Trips can be aggregated to whatever level of temporal detail/categorization is required by the network assignment model Deals naturally with “peak-spreading”, etc. Provides excellent detail for environmental impact analysis

Flexibility TASHA has been designed to be very flexible in terms of its development and its application. It has been developed using ordinary trip-based survey data for the GTA (but it could also exploit activity-based survey data). It can be used as a direct replacement for the first 3 stages in a 4- step system, or integrated within a full microsimulation model system. The data requirements for model development are no greater than other current models, including conventional trip-based models. Usable in a variety of contexts, and facilitates the evolution of the model system over time from aggregate to microsimulation.

Application in a conventional setting Pop & Emp by zone Standard 4-step zone-based inputs TASHA contains its own synthesis procedures to convert aggregate, zone-based inputs into disaggregated persons, etc. required for microsimulation Synthesize persons, hhlds & work/school locations TASHA Standard zone-based, static road & transit assignment Standard network assignment package (EMME, Vissum, etc.)

Application in a full microsimulation setting Base Year Census Data, Other Aggregate Data Application in a full microsimulation setting Synthesize Base Year Population, Employment, Dwellings, etc. ILUTE Evolutionary Engine For T = T0+1,T0+NT do: Demographic Update Building Stock Update Residential Housing Commercial Floorspace Firm/Job Location Update Household Composition Update Work/school Participation & Location Update Residential Location Update Auto Ownership Update Exogenous Inputs, Time T In-migration Policy changes … Travel Models Commercial Vehicle Movement Update Activity/Travel Update (TASHA) T0 = Base time point T = Current time point being simulated NT= Number of simulation time steps Dynamic Network Assignment Model (meso- or micro-scopic)

Current Status TASHA was developed using 1996 travel survey data for the GTHA. The activity scheduler has been validated against 2001 survey data. Interfaces with both EMME and MATSIM. TASHA has been used for several environmentally related studies in the GTHA. Has been applied in Montreal, London, Changzhou.

Environmental Modeling with TASHA TASHA has been connected with: EMME/2 road & transit network assignment model (link speeds & volumes by hour of day) MOBILE6.2C emissions model (link emissions by type by link by time of day) CALMET meteorological model CALPUFF dispersion model (pollutant concentrations by zone by time of day) Dynamic population exposure to pollution by zone by time of day!!

EXAMPLE INTERVENTIONS Transportation Policies (Road pricing, carbon taxes, transit investment, etc.) Land Use Policies Vehicle Technology Persons & Households Auto & Transit Travel Times/Costs TASHA Activity/Travel Scheduler Household Auto Ownership Model Transportation Network Model Activity Patterns & Trip Chains Trips By Mode, Vehicle Type & Time of Day VKT by Facility Type, etc. Vehicle Allocation Model Hot/Cold Soaks, Cold Starts, etc. Emissions Model Locations of People by Time of Day Exposure to Pollution Mobile Source Emissions Dispersion Model

Auto Emissions by location and time of day Link-based running emissions by time of day Zone-based soak emissions by time of day

Dispersion of Emission Concentrations

Zone NO2 Exposures

TASHA-MATSIM More recently TASHA has been linked with MATSIM, an agent-based micro/meso-scopic network simulator. MATSIM allows us to keep track of individual agents as they travel through the network so we can accumulate their emissions (and, eventually, their exposure to pollutants). It also provides us with rudimentary vehicle dynamics, allowing a more detailed calculation of vehicle emissions.

Implementing TASHA as GTAModel 4.0 TASHA assumes: Known work & school locations for all workers & students Known number of cars per household These models need to be added to TASHA (but would also have to be created for any V4.0 design) Will use: Doubly-constrained entropy PORPOW model (similar to GTAModel V2.0, V2.5 & GGH Model) Singly-constrained logit PORPOS model (similar to GGH Model) Simple logit household-based car ownership model

GTAModel V4.0 Pop & Emp by Zone Synthesize persons, households & jobs Household Car Ownership PORPOW PORPOS TASHA Activity generation Activity scheduling Tour-based model choice Auto allocation Ridesharing Emme Road & Transit Assignments by Time Period No Converged? Yes STOP

GTAModel V4.0 Pop & Emp by Zone Synthesize persons, households & jobs Household Car Ownership PORPOW PORPOS TASHA Activity generation Activity scheduling Tour-based model choice Auto allocation Ridesharing Includes shopping & other-purpose episode location choices (tour-context sensitive) Emme Road & Transit Assignments by Time Period No Converged? Yes STOP

GTAModel V4.0 Tasks Models to estimate: Household auto ownership level PORPOW PORPOS Shopping episode location choice Other-purpose episode location choice Activity episode generation rates Tour-based mode choice model (including car allocation & ridesharing models – jointly estimated)

Current Status 2012 24-hour Emme network will be ready “soon” (end of month?). Other data required being assembled (parking costs, fares, tolls, …). Will be ready to start estimating models as soon as TTS2012 available. Most models can be developed in parallel, expediting development time. 3-4 month development time to get the point where the model system can be operationally tested by City of Toronto (and anyone else wishing to do so).

Cloud Computing Earlier this week we successfully connected to the SOSCIP* cloud computing facility. This means that we can use the cloud to estimate the V4.0 models, greatly accelerating model estimation time. This will be particularly helpful for the tour-based mode choice models which are very computationally intensive to estimate. * Southern Ontario Super-Computing Innovation Platform. This is a joint venture of IBM Canada, Province of Ontario, Gov’t of Canada and most southern Ontario universities (led by UofT and Western) to provide the most powerful computing facilities in Canada to Ontario researchers and industry.

Risks TTS 2012 availability? Under-reporting of trips, especially in the PM peak. May need to calibrate the activity generation rates to improve fit to screenline counts by time of day.

Thank You!

GTAModel V4 – Data Requirements 2012 zonal parking costs 2012 station capacities and costs (already provided by Metrolinx) 2012 average adult transit fare for all agencies 407ETR 2012 tolls

Other Transit Lane Configurations Seaparated ROW ramp to terminal Surface ROW, exclusive lanes Surface ROW, mixed lanes c cs sl csx Proposed Solution c cs sl Vdf=60 or @xrow=1 or type=??? Alternative 1 c cs s u Alternative 2