Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 33

AGENT-BASED

MODELING
Güven Demirel

1
OUTLINE

 What is Agent-Based Modeling (ABM)


 Complexity Science & ABM
 Methodological Status of ABM
 Building Blocks of ABM
 Agent Definitions & Characteristics
 How to Model An Agent?

2
OUTLINE (ctd.)

 Different Approaches to Agents in ABM


 Multiple Agents & Interactions Among Agents
 Environment Types
 The Process of ABM Building
 An Example Model: SugarScape Model
 SD-ABM Links & Comparison

3
ABM Definitions
 “The main concern and interest of Agent-Based Modeling is the
experimental study and modeling of agents in a common world of
compound and unpredictable effects of a population of agents in a
common world.”
(Castelfranchi, C., Simulating with Cognitive Agents: The Importance of Cognitive Agents, MABS, 1998)

 “Agent-based modeling focuses on the interaction of rule-based


agents. The power and intelligence of real life system, for example, is
mimicked in the laboratory populated by agents that act in the model
space exactly according to rules observed in real life.”
(Scholl, H., J., Looking Across the Fence: Comparing Findings From SD Modeling Efforts with those of Other Modeling Techniques)

 “The aim of agent-based modeling is to look at global consequences of


individual or local interactions in a given space.”
(Reynolds, C., W., Individual-Based Modeling, 1999)

4
Complexity Science & ABM
 The modeling domain of ABM: complex systems.
 Complexity Science: the science of modeling and
analysis of complex systems.
 “Science of Emergence”
 Examples of emergence in sand pile, el Farol bar
and firefly models
 Unexpectedness and unpredictability
 Focus on nonlinear interactions
 Wholistic nature
(Ref.: Castelfranchi, C., Simulating with Cognitive Agents: The Importance of
Cognitive Emergence, MABS, 1998)

5
Complexity Science & ABM

Application Areas:
– Agent-Based Computational Economics
– Agent-Based Computational Demography
– Individual-Based Modeling in Ecology and Biology
– Political Science
– Anthropology
– Sociology
– Physics, etc.
(Ref.: Billari, C., F., et. al. Agent-Based Computational Modeling, 2006)
6
Complexity Science & ABM
 Research Institutes:
– Santa Fe
– New England
– Brookings

 Some Important Researchers in the Field:


– R. Axtell & J. Epstein (Growing Artificial Societies: Social
Science From the Bottom Up, 1996)
– R. Axelrod (The Complexity of Cooperation: Agent-Based
Models of Competition and Collaboration, 1997)
– D. Farmer
– T. Schelling (Micro Motives and Macro Behavior,1978)

7
Methodological Status of ABM

 ABM as a Social Simulation Methodology (as an


alternative to qualitative theories)
– Computer simulation of complex social processes
– Descriptive modeling
– More precise and formal definitions compared to qualitative social
sciences
– Used to test theoretical hypotheses
– Production of data and validation
– Toy-world character and policy analysis

(Ref.: Conte, R., MAS and Social Simulation, MABS, 1998


Gilbert, N., Terna, P. How to build and use agent-based models in social
science,1999)
8
Methodological Status of ABM

 ABM as a Micro Modeling Approach


– Constitution of Artificial Societies of Artificial Agents
– Models of Individual Agents and Interactions
– Capturing Heterogeneity
– Analysis of Emergent Phenomena
– Micro-Macro Link
– Application of Distributed Artificial Intelligence Techniques
– Is micro modeling always good?

(Ref.: Axtell, R., Epstein, J. Growing Artificial Societies: Social Science From Bottom
Up,1996)

9
Methodological Status of ABM

 ABM as an Alternative to Assumptions of Mathematical


Models
– An alternative to RAP
– Bounded rationality, not global optimization
– Rule-based formulation, where heuristic utility maximization is only one rule

 Deductive or Inductive
– Deductive in the sense that behavioral rules are given
– Inductive in the sense that these rules can be adaptive and if so should be
analyzed
– Inductive in the sense that behavior is analyzed

(Ref.: Gilbert, N., Terna, P. How to build and use agent-based models in social science,1999
Gilbert N. Quality, Quantity and Third Way, 2001)
10
Building Blocks of ABM

 Multiple Agents
– Individual Agent Layer
– Interactions Among Agents Layer

 Environment

11
What is an Agent?

 Perceive-Reason-Act Approach:
– perceive the environment, reason about its perceptions, and
act based on the reasoning (traditional AI).

 Thermostat, the simplest agent, people,


organizations, robots, etc.
12
What is an Agent?
 An (intelligent) agent perceives its environment via
sensors and acts rationally upon that environment with its
effectors.
 Rational mapping between percept sequences and
actions.
 Agents as computational processes that try to model the
capabilities of humans.
 Agent as an autonomous software component capable to
interact with environment and other agents and act
according to its agent program.

(Ref.: Wooldridge, M., Multiagent Systems,2001)

13
More on Agents
 Beliefs
 Goals
 Plans
 Agent Function& Agent Program
 Emergence of ABM from Distributed AI.
 Agents as computational processes implemented on
a computer that have:
– Autonomy
– Social ability
– Reactivity
– Proactivity

14
Agent Characteristics

 Autonomy (Freedom to act independently)


 Heterogeneity (different designs, no need to expose
internal structure)
 Adaptivity (Learning, being able to change decision
rules, goals, plans, etc. based on feedback from
system performance)
 Self-interested (work to reach own goals)

(Ref.: Wooldridge, M., Multiagent Systems,2001)

15
Types of Agents

 Simple Reflex Agents


 Model-Based Reflex Agents
 Goal-Based Agents
 Utility-Based Agents
 Learning-Agents

Note: This part is prepared mainly on Russell, S., Norvig, P.


Artificial Intelligence: A Modern Approach, 2003

16
Modeling of Simple Reflex Agents

17
Model-Based Reflex Agents

18
Goal-Based Agents

19
Utility-Based Agents

20
Learning Agents

21
Different Approaches to
Agents in ABM

 KISS Principle
 Complex Agents Perspective
 Agent type depends on model purpose and domain.
 Dominance of usage of collection of condition-action
rules in ABM.
 More use of AI algorithms, techniques. For the
design of adaptive (learning) agents:
– Learning by Induction
– Reinforcement Learning
– Learning through Artificial Neural Networks
– Learning by Genetic Algorithms
(Ref.: Billari, C., F., et. al. Agent-Based Computational Modeling, 2006)
22
Genetic Algorithm
 Evolutionary algorithm
 Biological analogy: Evolution by natural selection
 A population of individuals, each with some degree of “fitness”,
a metric explicitly defined by the modeler.
 The fittest individuals are “reproduced” by breeding them with
other fit individuals to produce offspring that share the features
from parents.
 Average fitness increases as population adapts to environment.
 Individuals of GA may be:
– Agents: Evolution of the agent population as a whole
– Agent Algorithms: Evolution of better algorithms, i.e. increases
utility.
(Ref.: Gilbert, N., Terna, P. How to build and use agent-based models in social
science,1999)
23
Multiple Agents & Interactions
Among Agents

 Communication is done via messaging.

 Interaction Types:
– Coordination
– Cooperation
– Negotiation
– Competition

24
Environment Types

 Fully vs. Partially Observable


 Deterministic vs. Stochastic
 Episodic vs. Sequential
 Static vs. Dynamic
 Discrete vs. Continuous

25
The Process of ABM Building
 Determination of whether the model to be constructed will be a
specific or generic model
 Determination of actual actors in the real system and software
agents are representations of a subset of these actors.
 Determination of model abstraction/aggregation level.
 Determination of what cognitively oriented computations the
agents can perform
 Selection of agent architecture
 Selection of software platform and model implementation

(Ref.: Doran, J. Agent design for Agent-based modeling,2006)

26
An Example Model

SUGARSCAPE

27
ABM vs. SD

 Two different perspectives in multiagent systems


modeling:
– Focus on individual, disaggregated agent actions
(micro-motives) modeling, the model overall behavior
emerges as a result of unfolding of the structure:
agents’ decisions and interactions. (Agent-Based
Modeling)
– Focus on macro, aggregated system variables, the
relations among the variables determine the model
behavior. (System Dynamics)

28
ABM vs. SD

SD
 Stock-flow structure
 Differential/difference equations, thus also called
equation-based modeling.
 Feedback theory, circular causality
ABM
 Agent Program
 Behavioral Rules & Interactions
 Adaptation and Learning
29
ABM vs. SD

Aggregation:
– SD is based on aggregation philosophy. SD makes an
abstraction from single events and individuals; and forms a
macro level modeling approach. Focus on system-level
variables, observables.
– ABM is a micro level modeling approach. ABM focuses on
the individual agents’ actions. ABM defines behavior at
individual level.

Policy Analysis:
– It is more natural to model multiagent systems by ABM for
policy analysis regarding individual agents.

30
ABM vs. SD

Adaptation:
– SD model structures are typically static.
– Agent decisions, reasoning, goals, etc. change by learning:
adaptation. As a result system structure may show
adaptation to changing conditions.

Heterogeneity:
– SD does differentiate among individuals.
– Each agent has its own design, need not expose its internal
structure to others in the system.

31
ABM vs. SD

Implementation:
– ABM may have more direct consequences for implementation
compared to SD, reference to direct inferences about individual
agent behaviors.
Validation:
– Structural Validation
– Behavioral Validation

(Ref.: Scholl, H., J., Agent-Based and System Dynamics Modeling: A Call for Cross study and Joint
Research, 2001
Scholl, H., J., Looking Across the Fence: Comparing Findings From SD Modeling Efforts with
those of Other Modeling Techniques
Parunak, H., V., D. et. al., Agent-Based Modeling vs. Equation-Based Modeling: A Case Study and
User’s Guide, MABS, 1998
Rahmandad, H., Heterogeneity and Network Structure in the Dynamics of Contagion: Comparing
Agent-Based and Differential Equation Models, ISDC, 2004
32 Demirel, G., Aggregated and Disaggregated Approaches to Multiple Agent Dynamics, ISDC,
2006 )
THANKS…

33

You might also like