Transcript Dynamic Spatial model - Centre for Advanced Spatial Analysis

Complexity Settlement Simulation
using CA model and GIS (proposal)
Kampanart Piyathamrongchai
University College London
Centre for Advanced Spatial Analysis
26 February 2003
Introduction
• Let’s introduce the work with this question ‘How growth
of a city and its surrounding hinterlands?’
• Human geographers have attempted to answer this
question by exploring many theories and models.
• Those are utilized as norm to explain the real urban
phenomena in several parts of the world.
• Evolutions of computer graphic and geographic
information sciences have resulted several new higher
potential ways to model the urban growth phenomena.
• It’s still not delimitated of imagination, how complex of
model can be applied by these technologies.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 2
Introduction
Cellular automata (CA) model
Dynamic spatial model (DSM)
t1
t2
Urban – Regional system
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 3
Introduction
• Integrating DSM to CA model makes possibility to define
complex spatio-temporal transition rules.
• Integrating DSM to the model makes possibility to use
powerful GIS functions in order to reveal spatio-temporal
behavior of urban and regional system.
• Integrating CA to DSM makes possibility to increasing the
temporal process of phenomena into spatial analysis.
• Focus more in urban-regional system to implement the
model is conceivable to set up cross-scale simulation
framework.
• Focus more in urban-regional system into the complex
model might be useful to improve some famous urban and
regional theories.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 4
Objective
To use of CA technique and GIS to model spatiotemporal complexity behavior of urban and regional
system
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 5
Scope of this study
• To integrate the CA simulation technique and dynamic
spatial model to reveal interaction of urban and regional
system.
• To explore variables or indicators in order to measure
the centrality and hierarchy of nodes (cities and towns in
a region) and find out the interaction and diffusion
pattern between entire nodes.
• To implement the complex settlement simulation
model.
• To recommend the model in order to use as decisionbased urban and regional management.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 6
Study Area
• The study area is
Bangkok municipality
and surrounding towns
and communities.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 7
Study Area
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 8
Study Area (Other choices)
• This might be reasonably to set up the simulation
model to other regions of Thailand because Bangkok city
is the apex metropolis. There are some difficulties to
utilize outcomes of the model to this city.
• It might be easier and efficiency to perform the model
and put it to use with the others that are taking after
Bangkok city.
• Other regions, can be potentially investigated, for
instance; ChiangMai and Phitsanulok in northern part,
KhonKaen and NakornRatchasima in northeastern part
etc.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 9
Work Plans
Month
1-5
6 - 12
13
-
24
25 - 28
29 - 36
100%
• Background theories review
• Literatures review
• Quantify centrality, hierarchy
and interaction of nodes
• CA Model and DSM
implementation review
• Data Collection
• Data manipulation
• Conceptual model
• Model implementation
• Data analysis and
programming
• Model Evaluation and Finalize
the user product
• Final writing up
0%
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 10
In progress
• Background Review : To catch the concepts, several papers,
notes, books and internet documents, concern to cellular
automata, dynamic spatial model, urban and regional system
theories, have being reviewed.
• Literatures Review : More precisely, the concerned papers
and other documents have being explored critically.
• This should be finished at the tenth month and the end of
first year of study respectively.
1-5
6 - 12
13
-
24
25 - 28
29 - 36
• Background theories review
• Literatures review
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 11
Conceptual framework (temporary)
Decision rule of
transition for
generating or not
generating a new cell
Dynamic
Spatial Model
Spatial information
from time t will be
used as base to
generate a new
situation in the
next time step
CA and Diffusion
model
Spatial information
in time t + 1
return to the
dynamic spatial
model
The CA model returns a new
situation at time t such as
size of urban suburban
Population Model
Spatial Interaction
Model
Nodal Model
Hierarchy Modification
Model
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 12
CA and Diffusion model
t1 -> t2
by
Conceptual framework (temporary)
• CA algorithm
• Mean information Field : MIF
Urban-region
Growth Map
Urban form State at t2
return
• Size of urban objects
• Compactness index
• Predictive population
• Gravity model
• Modified hierarchy of nodes
• Sphere of influence
by
Spatial interaction model
DSM
criteria
• Accessibility
• Transportation
• Potential surface
• Centrifugal&Centripetal
• Core-Periphery model
• Land use
• Urban-regional growth policy
Decision rule of transition
• etc.
Complexity Settlement Simulation using CA model and GIS
Dynamic maps
CASA : 26 Feb 2003 : 13
Node, Hierarchy and Centrality
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 14
CA and Diffusion model
T = 21
T=2
Spatial interaction
Predictive
population in
spatial database
Existing urban area
Traditional GIS in DSM framework
Complexity Settlement Simulation using CA model and GIS
Decision rule of transition in CA
and diffusion model
CASA : 26 Feb 2003 : 15
CA and Diffusion model
T = 32
T=3
Spatial interaction
Predictive
population in
spatial database
Existing urban area
Traditional GIS in DSM framework
Complexity Settlement Simulation using CA model and GIS
Decision rule of transition in CA
and diffusion model
CASA : 26 Feb 2003 : 16
Spatial Interaction
T=1
Sphere of influence?
Complexity Settlement Simulation using CA model and GIS
How and how much do both two nodes
interact spatially together?
Potential Surface?
CASA : 26 Feb 2003 : 17
Centrifugal and Centripetal forces
• Another interesting spatial interaction framework
• Two counteracting forces which are said to cause
changes in the pattern of land use in urban areas.
• The centripetal force causes centralization
attracting establishments to the central area where
they may benefit from the advantages of
accessibility and agglomeration economies.
• The centrifugal force causes decentralization and
urban sprawl as it pushes dwelling and businesses
away from congested, expensive inner city areas
towards the suburbs.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 18
Centrifugal and Centripetal forces
• To apply this two forces in this simulation, the
balance of two forces may be defined as levels.
• New cities tend to be activated by ‘centripetal
force’. This causes a new economic activities in new
towns.
• Unplanned and old cities tend to be stimulated by
‘centrifugal force’. This causes urban expansion and
sprawl.
• If the level of forces can be defined quantitatively,
this should be useful for making more reality.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 19
Dynamic Spatial model
Transition Rule
Static spatial
variables
Dynamic
spatial
variables
CA and Diffusion Model
Each time step
Spatial interaction model
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 20
Dynamic Spatial model
Static spatial
variables
Dynamic spatial variables
Complexity Settlement Simulation using CA model and GIS
Physical variables
Barrier and Impedance
Land policy
Other static vars.
………
Spatial interaction
Population&density
Existing urban area
Other predictive vars.
………
CASA : 26 Feb 2003 : 21
Data Base
CA and Diffusion
model
in Cell-based GIS
programming
Tabulated
data and
other
Documents
Multi-date
satellite
images
Cell-based
GIS modelling
GIS database
Field works
Maps
Complexity Settlement Simulation using CA model and GIS
Dynamic spatial model (DSM)
t1
t2
CASA : 26 Feb 2003 : 22
Next Steps to explore
• CA and Diffusion algorithms on GIS-based
programming
• How to implement spatial interaction model
• Which spatial interaction models are suitable for
this simulation model
• The criteria and weight definition in DSM
• How to visualize the model as dynamic maps and
urban –region growth maps
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 23
Expected Results
• The complex spatial model that can illustrate
simulated results both in urban and regional scale.
• Dynamic maps represent the growth of cities and
their neighbors.
• The framework of integration of CA model and
dynamic spatial model.
• A flexible simulation model that can be improved
as decision support system for urban and regional
management.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 24
Next Steps
• Defining specific study area and collecting
preliminary data.
• Criticizing more concerned documents in order to
fine-adjust the conceptual framework.
• Quantifying the centrality, hierarchy and interaction
of nodes (cities and towns) in urban and regional
scale.
• Preliminary model implementation for testing the
model.
Complexity Settlement Simulation using CA model and GIS
CASA : 26 Feb 2003 : 25