The Argos Freight Modeling Project

1
The Argos Freight Modeling Project
Presentation to SCAG Modeling Task Force 9/22/04
Genevieve Giuliano Peter Gordon School of Policy,
Planning and Development University of Southern
California
2
The Argos Research Project
Purpose Develop an intra-metropolitan freight
flow model using state of the art computer
science techniques
3
Argos Project Goals

• Contribute to knowledge in both information
  science and social science
• Address a significant government problem
• Conduct interdisciplinary research that bridges
  information science and social science

4
The Intra-metropolitan Freight Flow Estimation
Problem

• Models
• Goods movement as inter-regional trade problem
• Many good models of inter-regional freight flows
• High level of aggregation in geography,
  commodities
• Dont translate easily to intra-metropolitan
  context
• State-of-practice transportation planning models
• Conventional urban models do not take freight
  directly into account
• Traditional 4-step approach difficult to apply to
  freight
• Differences in behavioral foundations of travel
  demand
• Variation in supply/demand across industry
  sectors
• Changing industry dynamics

5
Problems, cont

• Data
• Extensive data requirements
• Flows by mode, commodity, value, origin
  destination
• Lack of disaggregate data
• US, state level data generally available county
  and smaller unit limited
• Fine geography does not coincide with fine sector
  data
• Proprietary data
• Private firms not required to release data
• Customs data processed by third parties
• Black box nature
• Survey data
• Infrequent, costly to update

6
An Ideal Model Would

• have solid behavioral foundation
• be multi-modal
• handle passenger and freight trip interactions
• minimize use of costly data
• be easily updated
• be transferable

7
Our Approach

• Use accessible, reliable data sources
• Available to public
• Regular updates
• Transparent computations
• Produced by reliable sources
• Use economic supply/demand as model foundation
• Use as much detail as possible
• Geography
• Economic sectors

8
Basic Concepts of Model

• Freight supply/demand in two basic categories
• Intra-regional supply/demand within SCAG region
• Inter-regional imports to, exports from,
  pass-through flows
• Intra-regional supply/demand
• Use IMPLAN I/O model for region
• Factor out imports/exports
• Inter-regional supply/demand
• Use import/export data sources (by mode)
• Identify import/export zones/nodes
• Use regional employment as control total

9
Main Data Sources
10
INTRA-REGION
IMPORTS/EXPORTS
WCUS 2000
SCAG SURVEY
I-O TABLE 2001
AIRPORT DATA
ITMS
SCAG 2000 EMP
LINKS TO IMPORT, EXPORT NODES
AGGREGATE SECTORS
CONVERSION TO COMPATIBLE SECTOR, /TON VALUES
CHECK
ATTRACTIONS PRODUCTIONS (, JOBS, by TAZ)
COMBINE INTER AND INTRA P A
FREIGHT FLOW MODEL STEPS IN GENERATING FREIGHT
SUPPLY/DEMAND
CHECK
AGGREGATE TO REGIONAL TOTALS
GO TO TRANSPORTATION MODEL
11
SOME DETAILS ON INTRA-METRO PORTION
SCAG 2000 EMP DATA Employment by sector, TAZ
IMPLAN 2001 I/O Regional transactions table
Check/adjust to sector, county totals
Apply sector IO coefficients to employment data
Generate supply, demand by TAZ, sector in jobs,
12
SOME DETAILS ON FREIGHT TRANSPORTATION MODEL
N x N matrix of freight Ps, As in tons
Factor out air, water, rail
Generate O-D matrix using gravity-type
distribution model
SCAG region transportation network with
equilibrium passenger traffic assignment
Convert O-D flows to PCEs
Traffic Assignment of freight PCEs
Equilibrium total traffic assignment
13
Workflow for Freight Estimation
14
Recent Work

• Update data sources to 2000, 2001
• Develop conversion table for sector codes (SIC,
  NAICS, SITC, etc)
• Seek better sources for ground imports/exports
• Seek better methods for /tons/jobs conversions
• Develop logical internal checks
• To learn more.http//www.isi.edu/argos