Urban Planning by Simulation of Population Growth

ocelotgiantAI and Robotics

Nov 7, 2013 (3 years and 5 months ago)

54 views

Urban Planning by Simulation of Population Growth



Cirano Iochpe

Flavio Rech Wagner

Marcia Aparecida da Silva Almeida

Guillermo Nudelmann Hess

André Dias Bastos


GEOINFO 2004

6th Brazilian Symposium on Geoinformatics

Outline


Related work


Introduction


The System










The beginning


Functions and functionalities


The architecture


Tools and technologies


The covering map


Modelling

Outline


Related work


Introduction


The System







Next Steps


Last considerations






The beginning


Functions and functionalities


The architecture


Tools and technologies


The covering map


Modelling

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

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


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

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


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

The InterSIG Simulation Subsystem
-

The Beginning


Visualizing influence zone of a public resource

The InterSIG Simulation Subsystem
-

The Beginning


Porto Alegre City Hall


Rules about public resources


Declivity < 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

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


The System Architecture

GML / SHP
Geographic
data
XML
rules
usuário
usuário
Geographic scenario
manager
(
wrapper
)
Geotools
API
Simulation
Covering
map
Simulador
Engine
Covering
Interface
WEB
Rules manager
(
Wrapper
)
Cellular
automata
Area
(
map
)
Neighbourhood
(
map
)
State
Transition
Rules
Files
(
shp
,
gml
)
Programa
Files
(Java)
Rules
files
(
xml
)
Metadata
DBMS
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






The covering map algorithm

School

Darn

Canal

Stream

Lake

River

Declivity

Generate buffer zone

Generate hidrology layer

Generate appropriated geographical zones

Generate influence zones

Sectors

Covering Map

The covering map algorithm

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



Finding a Model


Main approaches in dynamic modeling of
urban growth


Cellular automata


Heuristic methods


Neural networks


Finding a Model


Spatial Dynamic Modeling


Simulation of urban


land use changes


Claudia Almeida’s

(INPE) phD tesis


Empirical probabilistc


methods

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


Finding a Model


Bayesian networks








Qualitative aspect


Variables and their relationships (nodes and edges)


Quantitive aspect


Intensity degree of relationship between variables
(probabilities)


P(Xa
)

P(Xd|Xa
)

P(Xb|Xa
)

P(Xc|Xb,Xd
)

Finding a Model


Bayesian networks







Visit Asia?

(A)

Tuberculosis?

(T)

Lung cancer?

(C)

(T) or (C)?

(O)

Bronchisis?

(B)

Smooker?

(S)

X
-
ray+

(X)

Dispnesis

(D)

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

Conclusi
ons


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