AAG, NY, USA Feb. 28. 2012 COUPLING WITH SLEUTH, THE CVCA AND THE DG-ABC Dr. Elisabete A. Silva [email protected] Department of Land Economy University of

AAG, NY, USAFeb. 28. 2012

COUPLING WITH SLEUTH, THE CVCA AND THE DG-ABC

Dr. Elisabete A. [email protected]

Department of Land EconomyUniversity of Cambridge

Association of American Geographers, NY, USAFeb. 28, 2012


OUTLINE

1. The CVCA model

2. The people’s model

3. The DG-ABC model

4. Future: The importance of hybrid models in urban planning

5. Concluding remarks


URBAN MODELSLEUTH

ENVIRONMENTAL MODEL CVCA

Expert Decision Input

PEOPLES’ MODEL

Data Base

Urban ExcludedSlope(RAN-REN)TransportationHilshade

Past

Time

Future

SWOT


urban roads slope

excluded hilshade

1975 1976

.

.

.

1997

test mode

calibrate

forecast

coarse

fine

final

DNA

data acquisition

1998 1999

.

.

.

. 2025

metrics images

images metrics

reclass excluded

calculate LA metrics

apply LA strategies

1998

1999

-

SLEUTH results CVCA results

workshop’s morning

workshop’s afternoon

SWOT analysis

Map-SWOT & analysis-critic

reclassify SLEUTH’s input data

urban roads slope

excluded hilshade

Keep existent DNA?

yes?

no?

new DNA

forecast

SLEUTH Urban Model

CVCA EnvironmentalModel

Expert Inclusion‘people’s model’


urban roads slope

excluded hilshade

1975 1976

.

.

.

1997

test mode

calibrate

forecast

coarse

fine

final

DNA

data aquisition

1998 1999

.

.

.

. 2025

metrics images

images metrics

reclass excluded


apply LA strategies

1998

1999

CVCA - CA - Environmental Model

1. CVCA


Transition Rules:

Number of pixels (pixels with a probability of change to urban)

Action step

1. Protective 0 but NN > MNND

than add protective pixels around all outer patch and add protective pixels until arriving at closest neighbor

2.Defensive <=50% *,** than add defensive pixels to all outer patch cell where transition cell exists

3.Offensive >50% add offensive pixel to all outer patch cells and add offensive cells until nearest neighbor

4.Opportunistic 0 but NN = NNI (and no transition cell nearby) than link to nearest neighbor

5. Grow Goal or Result

D.Opportunistic

C. Offensive

B. Defensive

A. Protective

Core AreaBuffer Zone

Corridor

Supporting Landscape MatrixNon-Supporting Landscape Matrix

Desired network elements are identified and protected through planning policy and land use control in advance of negative landscape matrix changes.

Isolated core area in ‘non-supportive landscape matrix’ is subject to isolation from disturbance to corridors and to incremental reduction in size of the core area that can be protected through a new buffer zone.

Isolated core area is protected with a buffer zone and linked into a greenway network with corridors that are newly developed within a non-supportive landscape matrix context. The offensive strategy employs a range of tactics, including nature development, to achieve a desired landscape configuration.

Isolated core area is linked with an existing corridor, buffered, and anew supporting landscape matrix is developed. The opportunistic strategy takes advantage of unique circumstances that may only support some greenway uses, e.g. recreation.

Existing Landscape


Metric - AMP Value

Edges 14964

Area 24204

Num Clusters 708

MCS 34

MPS 275

LSI 7.7

MNND 1.5

Metric - AML Value

Edges 35171

Area 106460

Num Clusters 1134

MCS 93

MPS 577

LSI 9.9

MNND 1.6


CVCA Simulation


urban roads slope

excluded hilshade

1975 1976

.

.

.

1997

test mode

calibrate

forecast

coarse

fine

final

DNA

data acquisition

1998 1999

.

.

.

. 2025

metrics images

images metrics

reclass excluded


apply LA strategies

1998

1999

-

SLEUTH results CVCA results

workshop’s morning

workshop’s afternoon

SWOT analysis

Map-SWOT & analysis-critic

reclassify SLEUTH’s input data

urban roads slope

excluded hilshade

Keep existent DNA?

yes?

no?

new DNA

forecast

2. The People’s model


TWO MAP DRAWINGS RESULTING FROM THE WORKSHOP’S AFTERNOON


Strengths Votes % Weakness votes%

Opportunities Votes %

Threats votes

Transport system (road network, airport, harbor)

19.5 Mobility, accessibility and transport

32.6 Improve transportation system

17.6 Uncontrolled urban sprawl

29.9

Tourism and world heritage (Lisbon and Porto)

17.8 Lack of urban quality

17.0 Urban renewal 15.7 Natural risks (e.g. coastal, flooding, earthquake)

16.4

Capital city 13.0 Uncontrolled urban sprawl

11.3 Cultural tourism/events

11.8 Urban violence and drugs

14.2

SWOT RESULTS



3. DG-ABC MODEL

3.1 Concept model of DG-ABC model

Intelligent agents Cellular automata TPB model Genetic algorithm

Dynamics capturing

a-spatial dynamics spatial dynamics behavioural regulations

behavioural optimizations

Factors oriented

social-economic influences

infrastructures/ecosystems

behaviours of agents behaviours of agents

Level individual individual individual level high level

changes alter behaviours by GA and themselves

neighbourhoods navigation

N/A evolution by themselves

Data requirement

social-economics /policies quantifying

GIS data agent’s beliefs/ profile information

strategies/options

Integrated model


3. DG-ABC MODEL

3.2 DG-ABC model

1. Model Environment2. Heterogeneous agents3. CA (SLEUTH)4. Decision behaviors5. Interactions6. Synchronization

The key decision tables:

•The Resident agents’ utility table.•The developer agents’ development

application table. •The government agent’s approving

table. •Synchronization decision table.

spatial data

Source: Ning Wu and Elisabete A. Silva 2010a


3. DG-ABC MODEL

3.3 Theory of Planned Behavior

1 2 3( ) max{ , , .......... }nD f B Be Be Be Be

atraffic environment convenience tijEI a E b E c E

1 2 1 2Be W I W AbC W I W EI

a i iai p Ctd ctAp AW WA

1( ) /

na a ii ij ji neighbor

jSN M Inf N

1

ma a ai ik ki

kPBC Cb P

I

•A: the degree to which the performance of the behaviour ispositively or negatively valued.

•SN: an agent’s perception of social normative pressures, orrelevant others’ beliefs that the agent should (not) performsuch behaviour.

•PBC: an individual’s perceived ease or difficulty of performing the particular behaviour.

•I: an indication of a agent’s readiness to perform a givenbehaviour.

Behavioral Beliefs

Normative Beliefs

Subjective Norm

Perceived Behavioral Control

Control Beliefs

Intention

Actual Behavioral Control

Attitude toward the behavior

Behavior

TpB model (Icek Ajzen 2006)

Attitude toward the behavior

Subjective Norm

Perceived Behavioral Control

Intention

Behavior

Actual Behavioral Control

2 311 2 3( )highway citycenterroad

B D B DB Dt t ttrafficE w A e w A e w A e


Properties of resident agents

Properties of property developer agents

Properties of government developer agents


Spatial synchronization in the model

Temporal synchronization in the model


(a) run CA standby (b) run agents standby

(c) run integrated model (d) real urban data


4. FUTURE….HEXA-DPI



HEXA-DPI Data Structure and DME – Dynamic Model Environment


Key Issues:

-Historical expertise

-Interoperability.

-Spatial and A-Spatial dynamics

-MAUP- Transition matrix

- FUTURE…………. HEXA-DPI


REFERENCES

2012 (forthcoming) Surveying Models in Urban Land Studies. Journal of Planning Literature. ( with N. Wu)

2011 Cellular Automata Models and Agent Base Models for urban studies: from pixels, to cells, to Hexa-Dpi’s. In: Urban Remote Sensing: Monitoring, Synthesis and Modeling in the Urban Environment. Edited by: Dr. XiaojunYang. Wiley-Blackwell. pp. 323-345. ISBN: 978-0-470-74958-6

2010 A Planner’s Encounter with Complexity, (with G. de Roo) Ashgate Publishers Ltd, Aldershot (UK). 337 pages. ISBN: 978-1-

4094-0265-7. http://www.ashgate.com/isbn/9781409402657 2010 Waves of complexity. Theory, models, and practice. In: Roo, Gert de, and Elisabete A. Silva (2010), A Planner’s

Encounter with Complexity, Ashgate Publishers Ltd, Aldershot (UK). pp. 309-331.. ISBN: 978-1-4094-0265-7

2010 Complexity and CA, and application to metropolitan areas. In: Roo, Gert de, and Elisabete A. Silva (2010), A Planner’s Encounter with Complexity, Ashgate Publishers Ltd, Aldershot (UK). pp..187-207. ISBN: 978-1-4094-0265-7

2010 Artificial intelligence solutions for Urban Land Dynamics: A Review. (with N.Wu) Journal of Planning Literature. 2010 24: 246-265. ISSN: 0885-4122

2009 A Traffic Analysis Zone Definition: A New Methodology and Algorithm. (with LM Martinez and JM Viegas). Transportation. 36 (5): 581. 0049-4488 (Print) 1572-9435 (Online). DOI: 10.1007/s11116-009-9214-z

2009 (online 2008, print 2009), Modifiable Areal Unit Problem Effects on Traffic Analysis Zones Delineation. (with LM Martinez and JM Viegas) Environment and Planning B – Planning and Design. 36(4): 625-643 ( advance online publication, doi:10.1068/b34033. ISSN: 0265-8135 (print) ISSN: 1472-3417 (electronic - http://www.envplan.com/abstract.cgi?id=b34033).

2008 Strategies for Landscape Ecology in Metropolitan Planning: Applications Using Cellular Automata Models. (with J. Wileden, J. and J. Ahern), Progress in Planning, 70(4):133-177 - ISSN: 0305-9006

2007 Zoning Decisions in Transport Planning and their Impact on the Precision of Results. (with LM Martinez and JM Viegas)Transportation Research Record, 1994 (08): 58-65 - ISSN: 0361-1981

2005 Complexity, Emergence and Cellular Urban Models: Lessons Learned from Appling SLEUTH to two Portuguese Cities. (with K. Clarke) European Planning Studies, 13 (1): 93-115 – ISSN: 0965-4313

2004 The DNA of our Regions: artificial intelligence in regional planning. Futures, 36(10):1077-1094. – ISSN: 0016-3287

2002 Calibration of the SLEUTH Urban Growth Model for Lisbon and Porto, Portugal. Computers, (with K. Clarke) Environment and Urban Systems, 26 (6): 525-552 - ISSN: 0198-9715

Ajzen, I. 1985, From intentions to actions: A theory of planned behavior. In J. Kuhl & J. Beckman (Eds.), Action-control: From cognition to behavior (pp:11- 39). Heidelberg, Germany: Springer.

Documents

AAG, NY, USA Feb. 28. 2012 COUPLING WITH SLEUTH, THE CVCA AND THE DG-ABC Dr. Elisabete A. Silva [email protected] Department of Land Economy University of