Urban Planning by Simulation of Population Growth

1
Urban Planning by Simulation of Population
Growth
GEOINFO 2004 6th Brazilian Symposium on
Geoinformatics

• Cirano Iochpe
• Flavio Rech Wagner
• Marcia Aparecida da Silva Almeida
• Guillermo Nudelmann Hess
• André Dias Bastos

2
Outline

• Related work
• Introduction
• The System

• The beginning
• Functions and functionalities
• The architecture
• Tools and technologies
• The covering map
• Modelling

3
Outline

• Related work
• Introduction
• The System
• Next Steps
• Last considerations

• The beginning
• Functions and functionalities
• The architecture
• Tools and technologies
• The covering map
• Modelling

4
Introduction

• InterSIG Project
• Main goal to integrate a set of algorithms,
  techniques, tools, data models, and protocols
  into an Internet based system that supports both
  access and manipulation of geographic data

5
Introduction

• Simulation Subsystem of Geographic Scenarios
• Fase 3 of the InterSIG Project
• Main Goal to offer a web based simulation system
  that can be remotely used by municipalities to
  support urban planning activities
• Focused on urban growth
• Partners
• Data availability

6
Related Work

• A number of systems has been proposed to address
  urban growth simulation
• Most of them are not available on the Internet
• Most of them are dependent on specific GIS
  platforms and data formats
• UrbanSim
• Uplan

7
The InterSIG Simulation Subsystem - The Beginning

• Partnership Porto Alegre City Hall
• Project Planning the Future of the Lomba do
  Pinheiro District
• Availability of geographic data
• Hidrology, declivity, population,
• infrastructure
• Public resources schools,
• public health centers,
• kinder gardens, squares

8
The InterSIG Simulation Subsystem - The Beginning

• Partnership Porto Alegre City Hall
• Project Planning the Future of the Lomba do
  Pinheiro District
• Needs
• Visualize covering or influence area of a public
  resource
• Simulate inclusion of new resources
• Simulate increasing the population and its
  consequence to the
• covering area of public resources

9
The InterSIG Simulation Subsystem - The Beginning

• Visualizing influence zone of a public resource

10
The InterSIG Simulation Subsystem - The Beginning

• Porto Alegre City Hall
• Rules about public resources
• Declivity lt 25
• Each type of public resource has a specific range
  of influence given no geographic obstacles are
  provided
• Each instance of public resource has a maximum
  number of citizens which it can serve at any time

11
The System

• Functions and functionalities
• Accessible through the Web
• Upload of geographic scenarios by the user
• Upload of simulation rule sets by the user
• Generation of covering maps
• Simulation of the evolution of influence areas
  during a time interval

12
The System Architecture
13
Tools and Technologies

• Tools and technologies
• JSP/Servlets to interface
• GeoTools to handle geographic information
• GML, XML and Shapefiles to exchange data
• An owner simulator kernel in Java
• SVG to visualize maps

14
The covering map algorithm
School
Darn
Canal
Stream
Lake
River
Declivity
Generate hidrology layer
Generate buffer zone
Generate appropriated geographical zones
Sectors
Generate influence zones
Covering Map
15
The covering map algorithm
16
Modeling The Major Difficulty

• Find a urban growth population model
• Just rules are not enough to build the system
• Population is distributed following a
  growth/distribution model
• Each urban area can have a different model
• How to obtain a generic (basic) model?
• Dynamic modeling

17
Finding a Model

• Main approaches in dynamic modeling of urban
  growth
• Cellular automata
• Heuristic methods
• Neural networks

18
Finding a Model

• Spatial Dynamic Modeling
• Simulation of urban
• land use changes
• Claudia Almeidas
• (INPE) phD tesis
• Empirical probabilistc
• methods

19
Finding a Model

• Spatial Dynamic Modeling
• Simulation of urban land use changes
• Bayes theorem
• How about Bayesian Networks?
• At first glance, it can be used to obtain a model
  directly from a database

20
Finding a Model

• Bayesian networks
• Qualitative aspect
• Variables and their relationships (nodes and
  edges)
• Quantitive aspect
• Intensity degree of relationship between
  variables (probabilities)

21
Finding a Model

• Bayesian networks

Smooker? (S)
Visit Asia? (A)
Tuberculosis? (T)
Bronchisis? (B)
Lung cancer? (C)
(T) or (C)? (O)
X-ray (X)
Dispnesis (D)
22
Finding a Model Next steps

• Steps to build a Bayesian Network
• To obtain variables
• To obtain a causal relationship between variables
• From the database
• To obtain condicional probabilities tables
• Through bayesian learning methods

23
Conclusions

• To build a generic tool in terms of modeling is a
  quite dificult task
• To have the possibility of developing a new
  technique evolving dynamic modeling is quite
  interesting
• Next step is to continue investigating Bayesian
  Networks as the possible solution of our problem