UrbanSim: Integrated Land
Use, Transportation, and Environmental Modelling
|
|
|
Alan Borning
Dept. of Computer Science & Engineering |
|
Paul Waddell
Evans School of Public Affairs and
Dept. of Urban Design & Planning |
Other CSE People
Currently working on the Project
|
|
|
|
Graduate Students: |
|
Michael Noth |
|
L. Denise Pinnel |
|
Current & Former Undergraduate
Students: |
|
Michael Becke |
|
Nathan Freier |
|
Robin Cole |
Land Use/Transportation
Interactions
|
|
|
|
Patterns of land use and transportation
are critical in determining economic vitality, livability, and sustainability
of urban areas |
|
Land use and transportation interact
with each other and with the environment. |
|
Example: Suppose you were a Puget Sound
resident in the early 1960s. Three
alternative transportation plans are proposed. |
|
a freeway-oriented system |
|
a rail-oriented system |
|
do nothing |
|
|
Puget Sound Freeway Plans
from early 60s
Forward Thrust Light Rail
System
Puget Sound Example
continued
|
|
|
|
We might want to ask questions such as: |
|
What will the region look like in the
year 2000 under these different scenarios? |
|
How will people and goods move around? |
|
What will be the density and character
of development? |
|
How much open space will there be? |
|
What will be the environmental impacts
on air and water quality? |
Applications to Date
|
|
|
Eugene/Springfield, Oregon (including
historical validation — started with 1980 data and predicted state in 1995) |
|
Salt Lake City |
|
Honolulu |
|
now working on Puget Sound
(substantially larger area than the others) |
Software Status
|
|
|
Written in the Java programming
language (about 200,000 lines, plus 100,000 more lines of Java that is
automatically generated) |
|
Software is under the GNU Public
License (so it’s free, and freely redistributable) |
|
We have made an initial release of the
system, with several hundred downloads |
Software Architecture –
Overview
|
|
|
|
UrbanSim is composed of interacting
models that simulate particular aspects of the urban environment |
|
Demographics (people moving in and out
of the area, births and deaths) |
|
Residential location choice |
|
Business location choice |
|
Transportation |
|
… |
|
|
Slide 10
Spatial Representation
|
|
|
|
Geographic information is of the
essence! |
|
UrbanSim currently represents a region
using a 2-dimensional grid. |
|
Analogy: Game of Life grid |
|
Each grid cell is 150 meters square |
|
Grid cell attributes include: |
|
Building type, square feet |
|
Number of residents |
|
Number of employees (for businesses) |
|
etc |
Hawaii -- Housing Density
Hawaii -- Grid Cells
Time
|
|
|
|
As with most simulations, UrbanSim
keeps track of simulated time |
|
There is a global variable currentYear |
|
Analogy: “stepCount” in Game of Life |
|
Typical simulation is run for 20 years |
|
Most models run once each simulated
year |
Temporal Representation
The Model Coordinator
|
|
|
|
Handles initialization of models |
|
Coordinates execution of all models |
|
Notifies models of updates |
|
For example, the Land Developer model
might build some new houses. |
|
The Model Coordinator notifies the
Residential Location Choice model that new housing is available. |
Business Location Model
Nested Logit Structure
Latent Demand
Market Price Adjustment