Version September 11, 2006, 11:20 EST Agent-based …rlatimer/mongolia/NSFPoster_v4e.pdf ·...
1
How did Mongol s ociety develop the largest territiorial empire in the world history with only minor social complexity? 3000 B.C. 2000 B.C. 1000 B.C. A.D. 1000 A.D. 2000 Andean IS West Asian IS East Asian 1 IS Euro- W.Asian IS Global Modern Int’l System Chronological time τ 15th cent. A.D. fusion 16th cent.B.C. expansion Mesoamerican IS 4000 B.C. 5000 B.C. 6000 B.C. 19th cent. A.D. fusion 0 East Asian 2 IS MOTIVATION THEORY AND HYPOTHESES RESEARCH METHODOLOGY EXPECTED PROJECT RESULTS: 2006-09 PROJECT-RELATED PUBLICATIONS TARGET HISTORICAL SYSTEMS Smithsonian Institution G polity exists S situational change polity dealing w/issue post-issue polity Top-level UML state diagram for the fast canonical process.The referent system is a generic polity, subject to occasional situational changes (public issue occurrence) in its internal or external environment. outcomes occur Xiongnu in -215 Qin attack from S. Xiongnu defend post-attack Xiongnu outcomes occur High-level diagram for an instance of the fast process: The Xiongnu are attacked by the Chinese. They defend themselves and, as a result, the Xiongnu polity—its state— is changed. In this instance the state became more complex. Agents Societies SocialWorld InnerAsia Climate Enviroment Landscape Affects Affects Affects 1 1..n 1 1..n Interact Computational objects that compose the InnerAsia model Version September 11, 2006, 11:20 EST Main scientific puzzles addressed: • How do societies evolve or fail when challenged by threats or opportunities? • How does success or failure in collective action affect social complexity? • How did polities rise and fall in Inner Asia since ca. 3000 BC? Main gaps in needed scientific knowledge: • Lack of theory: Needed to advance our understanding of long-term societal dynamics • Lack of data: Needed for knowing the history of when, where, and how polities evolved • Lack of models: Needed for developing validated/calibrated agent-based computational models with sufficiently complex and shiſting environments Polities evolve by a process of “canonical variation” (Cioffi -Revilla 2005), whereby a society increases or decreases its political complexity (overall governance capacity) depending on success or failure in meeting collective action challenges (threats or opportunities). Sociopolitical evolution by canonical variation occurs on two distinct time-scales: •A fast process determines whether a society is successful or not in meeting collective action needs •A slow process determines whether a society develops or declines depending on the cumulative effects of numerous and oſten parallel fast processes Inner Asia (red oval) is a region that links the sociopolitically “pristine” regions of East Asia and West Asia aſter 3000 BC. Source: Cioffi-Revilla (2006). What Is MASON? Multi-Agent Simulator of Networks and Neighborhoods Main features: • Fast and 100% Java • Separable computation - visualization • Guaranteed reproducible results • Free and open-source hp://cs.gmu.edu/∼eclab/projects/mason/ Acknowledgements: Acknowledgements: This project is funded by the U.S. National Science Foundation, This project is funded by the U.S. National Science Foundation, Human and Social Dynamics Program, grant no. Human and Social Dynamics Program, grant no. 0527471. 1. New scientific concepts, theories, and methods for understanding long-term societal dynamics in complex and shiſting environments. 2. New databases on geographic, social, economic, political, military, technological aributes of Inner Asian societies, starting from ca. 3000 BC, including information on their physical environments. 3. New computational models—multi-agent systems or agent-based models—of societal evolution and adaptation, using the MASON system and evolutionary computation (see Research Methodology). Cioffi-Revilla, Claudio. 2005. A Canonical Theory of Origins and Development of Social Complexity. Journal of Mathematical Sociology 29 (April-June):133–153. Cioffi-Revilla, C., S. Luke, D. C. Parker, J. D. Rogers, W. W. Fitzhugh, W. Honeychurch, B. Frohlich, P. DePriest, and N. Bazarsad. 2006. Agent-based dynamics of social complexity: Modeling Adaptive Behavior and Long-Term Change in Inner Asia. Proceedings of the North American Association for Computational Social and Organizational Sciences NAACSOS 2006, June 22-23, Notre Dame, IN, USA. Cioffi-Revilla, C., S. Luke, D. C. Parker, J. D. Rogers, W. W. Fitzhugh, W. Honeychurch, B. Frohlich, P. DePriest, and N. Bazarsad. 2006. Agent-based dynamics of social complexity: Modeling Adaptive Behavior and Long-Term Change in Inner Asia. Proceedings of 20th World Congress of the International Political Science Association IPSA, July 9- 12, Fukuoka, Japan. Cioffi-Revilla, C., S. Luke, D. C. Parker, J. D. Rogers, W. W. Fitzhugh, W. Honeychurch, B. Frohlich, P. DePriest, and N. Bazarsad.. 2006. Agent-based dynamics of social complexity: Modeling Adaptive Behavior and Long-Term Change in Inner Asia. Paper presented in the First World Congress of Social Simulation, August 21-25, Kyoto, Japan. Powell, Eric A. 2006. Mysterious Mongolia. Archaeology 59 (1). Rogers, J. Daniel, Erdenebat Ulambayar, and Mahew Gallon. 2004. Urban Centers and State Development in Eastern Inner Asia. Washington, DC: Department of Anthropology, National Museum of Natural History, Smithsonian Institution. Agent-based Modeling and Simulation of Adaptation and Long-Term Change in Inner Asia (v.0.1) C. CIOFFI-REVILLA, S. LUKE, D.C. PARKER, M. TSVETOVAT, George Mason University J.D. ROGERS, W.W. FITZHUGH, W. HONEYCHURCH, B. FROHLICH, P. DePRIEST, Smithsonian NMNH R. LATIMER, T. Jefferson High School of Science and C. AMARTUVSHIN, Mongolian Academy of Sciences Student Researchers: G.C. Balan, R. Casstevens, C. Eltrich, J. Harrison, M. Latek, M.M. Rizi, S. Wilcox Main methods employed by our team: • Interdisciplinary collaboration • Field research at archaeological sites (see red dots on the top right picture) • Archival data collection • GIS, including social and historical • UML for designing models • Agent-based modeling using MASON • Evolutionary computation using ECJ Why is the fast process called "canonical"? • Because the same common theme in the event sequence G-C-N-U-S re-occurs again and again, throughout history, but with important variations caused by success and failure. How does the fast process with canonical variations "feed into" the slow process? • By a mechanism of "credit assignment", whereby different outcomes of the fast process produce positive and negative externalities for governance capacity E group endures w/o polity X group destroyed not knowing C X group is destroyed knowing C X group escapes or is destroyed A political complexity acrues G group exists C situational change occurs ~ C N collective action need or opportunity is recognized U locals undertake CA S CA succeeds ~ S ~ U locals don't undertake CA ~ N CA is not recognized political outcome (sample) space "back" to the next time a situational change occurs, C', for a new canonical iteration that will most probably differ in the details (sub-model) but not in this main theme "Fast" branching process for political evolution by canonical variation producing diverse outcomes for society undergoing challenging situational changes (threats or opportunities) Informal "canonical" theory Formal model of canonical theory Agent-based model code (PoliGenModel) Empirical data files: -archaeological -ethnographic -environmental testable predictions Prehistoric and historic record on emergent social complexity (explanandum) implementation formalization conceptualization feedback coding publish data in websites: • VDC • ICPSR • LTC • others Simulation environment (MASON) Post MASON and ABM models as open-source Formal and empirical components in the canonical theory of emergent social complexity.
Transcript of Version September 11, 2006, 11:20 EST Agent-based …rlatimer/mongolia/NSFPoster_v4e.pdf ·...
How did Mongol s ociety develop the largest territiorial empire in the world history with only minor social complexity?
3000 B.C.
2000 B.C.
1000 B.C.
A.D. 1000
A.D. 2000
Andean IS
West Asian
IS
East Asian 1 IS
Euro-W.Asian
IS
Global Modern
Int’l System
Chronological time τ
15th cent. A.D. fusion
16th cent.B.C. expansion
Mesoamerican IS
4000 B.C.
5000 B.C.
6000 B.C.
19th cent. A.D. fusion
0
East Asian 2 IS
Pleogenic model of political origins and evolution
MOTIVATION
THEORY AN D HYPOTHESES
RESEARCH METHODOLOGY EXPECTED PROJECT RESULTS: 2006-09
PROJECT-RELATED PUBLICATIONS
TARGET HISTORICAL SYSTEMS
Smithsonian Institution
Gpolityexists
S situationalchange
politydealing w/issue
post-issue polity
Top-level UML state diagram for the fast canonical process.The referent system is a generic polity, subject tooccasional situational changes (public issue occurrence) in its internal or external environment.
outcomesoccur
Xiongnu in-215
Qin attackfrom S.
Xiongnudefend
post-attackXiongnu
outcomesoccur
High-level diagram for an instance of the fast process: The Xiongnu are attacked by the Chinese. Theydefend themselves and, as a result, the Xiongnu polity—its state— is changed. In this instance the state became
more complex.
Agents
Societies
SocialWorld
InnerAsia
Climate
Enviroment
Landscape
Affects
Affects
Affects
1
1..n
1
1..n
Interact
Computational objects that compose the InnerAsia model
Version September 11, 2006, 11:20 EST
Main scientific puzzles addressed:• How do societies evolve or fail when challenged by threats or opportunities?• How does success or failure in collective action affect social complexity?• How did polities rise and fall in Inner Asia since ca. 3000 BC?
Main gaps in needed scientific knowledge:• Lack of theory: Needed to advance our understanding of long-term societal dynamics• Lack of data: Needed for knowing the history of when, where, and how polities evolved• Lack of models: Needed for developing validated/calibrated agent-based computational models with sufficiently complex and shifting environments
Polities evolve by a process of “canonical variation” (Cioffi-Revilla 2005), whereby a society increases or decreases its political complexity (overall governance capacity) depending on success or failure in meeting collective action challenges (threats or opportunities).
Sociopolitical evolution by canonical variation occurs on two distinct time-scales:• A fast process determines whether a society is successful or not in meeting collective action needs• A slow process determines whether a society develops or declines depending on the cumulative effects of numerous and often parallel fast processes
Inner Asia (red oval) is a region that links the sociopolitically “pristine” regions of East Asia and West Asia after 3000 BC. Source: Cioffi-Revilla (2006).
What Is MASON?Multi-Agent Simulator of Networks and
NeighborhoodsMain features:• Fast and 100% Java• Separable computation - visualization• Guaranteed reproducible results• Free and open-sourcehttp://cs.gmu.edu/∼eclab/projects/mason/
Acknowledgements: Acknowledgements: This project is funded by the U.S. National Science Foundation, This project is funded by the U.S. National Science Foundation, Human and Social Dynamics Program, grant no. Human and Social Dynamics Program, grant no. 0527471..
1. New scientific concepts, theories, and methods for understanding long-term societal dynamics in complex and shifting environments.
2. New databases on geographic, social, economic, political, military, technological attributes of Inner Asian societies, starting from ca. 3000 BC, including information on their physical environments.
3. New computational models—multi-agent systems or agent-based models—of societal evolution and adaptation, using the MASON system and evolutionary computation (see Research Methodology).
Cioffi-Revilla, Claudio. 2005. A Canonical Theory of Origins and Development of Social Complexity. Journal of Mathematical Sociology 29 (April-June):133–153.Cioffi-Revilla, C., S. Luke, D. C. Parker, J. D. Rogers, W. W. Fitzhugh, W. Honeychurch, B. Frohlich, P. DePriest, and N. Bazarsad. 2006. Agent-based dynamics of social complexity: Modeling Adaptive Behavior and Long-Term Change in Inner Asia. Proceedings of the North American Association for Computational Social and Organizational Sciences NAACSOS 2006, June 22-23, Notre Dame, IN, USA.Cioffi-Revilla, C., S. Luke, D. C. Parker, J. D. Rogers, W. W. Fitzhugh, W. Honeychurch, B. Frohlich, P. DePriest, and N. Bazarsad. 2006. Agent-based dynamics of social complexity: Modeling Adaptive Behavior and Long-Term Change in Inner Asia. Proceedings of 20th World Congress of the International Political Science Association IPSA, July 9-12, Fukuoka, Japan.Cioffi-Revilla, C., S. Luke, D. C. Parker, J. D. Rogers, W. W. Fitzhugh, W. Honeychurch, B. Frohlich, P. DePriest, and N. Bazarsad.. 2006. Agent-based dynamics of social complexity: Modeling Adaptive Behavior and Long-Term Change in Inner Asia. Paper presented in the First World Congress of Social Simulation, August 21-25, Kyoto, Japan.Powell, Eric A. 2006. Mysterious Mongolia. Archaeology 59 (1).Rogers, J. Daniel, Erdenebat Ulambayar, and Matthew Gallon. 2004. Urban Centers and State Development in Eastern Inner Asia. Washington, DC: Department of Anthropology, National Museum of Natural History, Smithsonian Institution.
Agent-based Modeling and Simulation of Adaptation and Long-Term Change in Inner Asia (v.0.1)
C. CIOFFI-REVILLA, S. LUKE, D.C. PARKER, M. TSVETOVAT, George Mason UniversityJ.D. ROGERS, W.W. FITZHUGH, W. HONEYCHURCH, B. FROHLICH, P. DePRIEST, Smithsonian NMNH
R. LATIMER, T. Jefferson High School of Science and C. AMARTUVSHIN, Mongolian Academy of SciencesStudent Researchers: G.C. Balan, R. Casstevens, C. Eltrich, J. Harrison, M. Latek, M.M. Rizi, S. Wilcox
Main methods employed by our team:• Interdisciplinary collaboration• Field research at archaeological sites (see red dots on the top right picture)• Archival data collection• GIS, including social and historical• UML for designing models• Agent-based modeling using MASON• Evolutionary computation using ECJ
Why is the fast process called "canonical"?• Because the same common theme in the event sequence G-C-N-U-S re-occurs again and again, throughout history, but with important variations caused by success and failure.How does the fast process with canonical variations "feed into" the slow process?• By a mechanism of "credit assignment", whereby different outcomes of the fast process produce positive and negative externalities for governance capacity
Egroup endures
w/o polity
Xgroup destroyed not
knowing C
Xgroup is destroyed
knowing C
Xgroup escapes or is
destroyed
Apolitical complexity
acrues
Ggroupexists
Csituational change occurs
~ C
Ncollective action need
or opportunity is recognized
Ulocals undertake CA
SCA succeeds
~ S
~ Ulocals don't undertake CA
~ NCA is not recognized
politicaloutcome(sample)
space"back" to the next time a
situational change occurs,C', for a new canonical iteration
that will most probablydiffer in the details (sub-model)
but not in this main theme
"Fast" branching process for political evolution by canonical variation producing diverse outcomes forsociety undergoing challenging situational changes (threats or opportunities)
Informal"canonical" theory
Formal model ofcanonical theory
Agent-based model code(PoliGenModel)
Empirical data files:-archaeological-ethnographic-environmental
testablepredictions
Prehistoric and historic record onemergent social complexity
(explanandum)
implementationformalization
conceptualization
feedback
coding
publish data inwebsites:
• VDC• ICPSR
• LTC• others
Simulation environment(MASON)
Post MASON and ABMmodels as open-source
Formal and empirical components in the canonical theory of emergent social complexity.
![Modeling Multi-Agent Systems using UML€¦ · UML-RT (UML-Real Time) [Selic and Rumbaugh 1998] as an ADL for Tropos. Part of the UML-RT concepts have been incorporated as architecture](https://static.fdocument.org/doc/165x107/5f0e369a7e708231d43e27e2/modeling-multi-agent-systems-using-uml-uml-rt-uml-real-time-selic-and-rumbaugh.jpg)
