BALTIMORE%20METROPOLITAN%20COUNCIL%20MODEL%20ENHANCEMENTS%20FOR%20THE%20RED%20LINE%20PROJECT

1
BALTIMORE METROPOLITAN COUNCIL MODEL ENHANCEMENTS
FOR THE RED LINE PROJECT

• AMPO TRAVEL MODEL
• WORK GROUP
• March 20, 2006

2
BMC Enhanced ModelBackground

• Update to the Baltimore region travel demand
  model
• MTA provided funding
• Developed by William G. Allen with assistance
  from BMC staff
• Motivation/Goal
• Revised mode choice model to meet FTA technical
  analysis of user benefits utilizing SUMMIT
• Analysis of proposed Red Line to request New
  Starts funding
• Develop one regional model used for long/short
  range regional planning, regional conformity
  analysis, and FTA new starts analysis

3
BMC Modeling Area
BMC Model Area
MWCOG Jurisdictions in BMC Model
4
Baltimore Region Highway and Rail Transit Systems
5
BMC Enhanced ModelObjectives

• User friendly logical input file names
• Reduce off-model data preparation
• Reduce number of input files
• Incorporation of 2001 NHTS data
• Meet Requirements of FTA
• Remove cliff functions
• Use same coefficients for transit and auto time
• Eliminate geographic bias constants
• Produce input files for SUMMIT

6
BMC Enhanced ModelStatus

• Draft functioning model as of February 2006
• Conducting sensitivity test designed to evaluate
  estimated changes in demand
• BMC staff developing a further understanding of
  the model through application, documentation, and
  development of user guide/local training session
• Extensive quality control checking of all model
  aspects
• Model runs in 9 to 10 hours (previous model 6 to
  7 hours)

7
Trip Generation Features

• Demographic models
• Households by size (5)/income (4)
• Urban Size (Baltimore City, DC, Montgomery, PG)
• Suburban Size (All Others)
• Urban Income (Baltimore City and DC)
• Suburban Income (All Others)
• Workers (4)/income (4)
• Workers for Washington region must be estimated
  based on household size and income from PUMS
• Income split is 10/15/20/55 of households
  (lowest to highest)

8
Trip Generation Features (Contd)

• 10 trip purposes
• Home-Based Work, Shop, Other, and School
• Non-Home-Based JTW, JAW, and OBO
• Truck CV, MTK, and HTK
• Nine area types based on household and employment
  density
• Trip productions by rates for different HH
  size/income and worker/income combinations
• Trip attractions by regression equation, adjusted
  by area type

9
Trip Generation Features (Contd)

• Home-based I-I trip attractions normalized to
  match productions
• Non-home based trip attractions normalized to
  productions, then productions set equal to
  attractions
• External trip generation based on
  counts/projections and distance between zone and
  closest external
• X-X trips based on base year trip table
• Attraction split model by income

10
2000 Density Codes
11
2030 Density Codes
12
Trip Distribution Features

• Composite Time
• Highway, transit (2 lowest income classes, HBW,
  HBS, HBO), toll
• Simulated congested time used for HBW, JTW,
  free-flow time for other purposes during first
  pass
• Two loops of speed feedback for HBW and JTW trips
• K-factors replace barrier penalties
• Truck deltas applied after distribution
• Truck deltas now additive, not multiplicative

13
K-Factor Map
14
Mode Choice Features

• New transit access pre-process
• TP, ArcInfo, Microsoft Access
• Create up to 9 walk links per mode, unlimited
  drive links within 30 minutes
• Create ½ mile buffer around local transit lines
  and express bus and rail stops
• Identify walk access share of each zone within
  buffer
• Determine average walk time for bus, rail,
  commuter rail, peak and off-peak
• Select links from those actually used in
  pathbuilding limited in number and distance/time

15
Mode Choice Features

• New nested logit mode choice model
• Same structure for all non-school purposes
• Model for school trips includes a top-level
  school bus nest
• Parking cost model
• Three levels of auto occupancy
• Three transit modes
• Two access modes

16
Mode Choice Features (Continued)

• Peak and off-peak transit networks
• Freeze-dry Washington mode choice
• Dummy links for DC transit access/egress

17
Mode Choice Structure
18
Mode Choice StructureSchool
19
Assignment Features

• Fixed weights based on future equilibrium
  assignment
• First two feedback loops assign AM peak trips
  (630-930 a.m.)
• Final assignment for four time periods
• AM peak (630-930 a.m.)
• Midday (930 a.m.-330 p.m.)
• PM peak (330-630 p.m.)
• Night (630 p.m.-630 a.m.

20
Assignment Features (Continued)

• Number of iterations
• AM peak 14
• Midday 8
• PM peak 14
• Night 6

21
Next Steps

• Test Alternatives
• Run SUMMIT
• Continue checking results
• Submit to FTA
• Document