Uploaded by Matteo Rucco

02-AgentModelling

advertisement
Agent-Based
Modelling and Simulations
Agent-Based Modelling
Agostino Poggi
Complex Systems
♦ Is a highly structured system, which shows
structure with variations
♦ Its evolution is very sensitive to initial conditions or
to small perturbations
 Number of independent interacting components is large
 There are multiple pathways by which the system can
evolve
♦ Is difficult to understand and verify by design or
function or both
♦ There are multiple interactions between many
different components
♦ Constantly evolves over time
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
2
Numeric Simulation Limits
♦ Equational models have a large number of
parameters
♦ Different theories must be used in biology,
sociology, economy, ...
♦ Difficulty of the micro/macro transition and
difficulty to represent different levels
♦ No representation of the behaviors, but only their
overall result (e.g., number of individuals, amount
of food, …)
♦ Doesn’t account for the emergence of spatial and
time structures (e.g. fish schools or flocks of birds,
columns of ants, ...)
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
3
Agent-Based Modelling
♦ Is an approach to modeling systems based on
autonomous and interacting agents
♦ Is a bottom-up process
♦ Defines emergent phenomena from microbehaviors
♦ Supports both optimization models and
investigation of a dynamic process
♦ Succeeds where centralized planning and
optimization models fail
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
4
Modeling Objectives (1/2)
♦ To understand some systems
 In detail (quantitatively)
 Qualitatively (the relationships between variables of
interest)
 To sharpen our intuitions
♦ To forecast or backcast some systems
 Behaviors of participants in the system
 System states (micro, meso or macro-level)
♦ To support intervention in some systems
 To advise participants on their strategies
 To advise owners (e.g., policy-makers) on their
management
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
5
Modeling Objectives (2/2)
♦ To create some reality (i.e., models)
 Black-Scholes options-pricing theory
 Game theory for nuclear weapons doctrines
…
♦ To enable co-ordination between stakeholders (i.e.,
models as co-ordination artefacts)
 Forecasting models in hedge funds
 Corporate strategy modeling
 Large-scale public policy modeling (national macroeconomic models, models of climate change,
communicable disease models)
…
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
6
Agent Types
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
7
Agent Schools
♦ Artificial intelligence
 Agents as autonomous entities solving problems
♦ Multi-agent systems
 Distributed control of systems
♦ Agent-based modeling (and simulation)
 Simulating (real world) phenomena
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
8
Agent Features
♦ Encapsulated
 Clearly identifiable, with well-defined boundaries and
interfaces
♦ Situated in a particular environment
 Receives input through sensors and acts through effectors
♦ Capable of flexible action
 Responds to changes and acts in anticipation
♦ Autonomous
 Has control both over its internal state and over own
behavior, reacts to environmental change and proactively
changes its behavior
♦ Designed to meet objectives
 Attempts to fulfill a purpose, solve a problem, or achieve
goals
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
9
Agent Model (1/2)
♦ An agent is a persistent thing which has some state
and which interacts with other agents, mutually
modifying each other’s states
♦ The components of an agent-based model are
 A collection of agents and their states
 Rules governing the interactions of the agents
 Environment within which they live
♦ Interaction among agents is the central point of the
simulation
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
10
Agent Model (2/2)
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
11
Environment
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
12
Agent Interaction
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
13
Behavior Ingredients
♦ Rule based
 Nested if-then-else structures
♦ Multi criteria decision making
 Options and weights
♦ Inference engines
 Expert systems, facts (states) and decision heuristics
♦ Machine learning
 Neural networks, deep learning, Bayesian statistics and
pattern recognition
♦ Evolutionary computing
 Find a optimal solution in large solution space (genetic
algorithms)
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
14
Simulation
♦ Agent models are used as substitutes for another
system
♦ Simulations mostly use virtual time
♦ Agents live a in a simulated environment
 Social space
 Virtual 2D/3D space
♦ Time and environment are controllable by the
modeler
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
15
Time
♦ Simulations take place in discrete time
♦ Time progresses in ticks
♦ Between two ticks, everything is assumed to
happen in the same time, attempting to simulate
the parallelism in real world
♦ As computers are serial processing machines, the
order of iterations among agents is very important
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
16
Behavior Vs Goal Oriented Models
♦ Behavior-oriented modeling
 Agents are described by modeling their behaviors
♦ Goal-oriented modeling
 Agents are capable of planning and the modeler
described their goal
♦ Choice of modeling strategy strongly depends on
application context
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
17
Behavior Oriented Models
♦ Modeler describes agent status and dynamics
♦ Examples of formalisms are activity graphs,
crisp/fuzzy rules, constraints, ...
♦ Reactions to perceptions/status changes are
defined by the modeler
♦ Can easily accommodate reinforcement learning
and evolutionary concepts
♦ Agents’ goal(s) are treated implicitly
♦ Very intuitive mapping with simple biological
systems (e.g., insects)
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
18
Goal Oriented Models
♦ Modeler identifies goals of the agents
♦ Agents select a goal and execute actions as a
consequence
♦ Reactions are not predefined, but goal dependent
♦ Explicit treatment of goals in the agent behavior,
but
 Execution of goal dependent actions can be error-prone
 Leads to significantly more complex model (see BeliefDesire-Intention agent models)
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
19
Development & Use
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
20
Advantages
♦ Allows appropriate modeling capabilities in a number
of important disciplines
 Social science, biology, software development, …
♦ Allows to simulate systems that are particularly
difficult to treat with traditional approaches
 Emergent phenomena, models with variable structure
♦ Can afford more detail in models
 More realism and micro-validity
♦ Provides an intuitive way of modeling
 Facilitates communication with other fields and enables
more researchers to use simulation
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
21
Applicability (1/2)
♦ When there are decisions and behaviors that can
be well-defined
♦ When it is important that agents adapt and change
their behaviors
♦ When it is important that agents have a dynamic
relationship with other agents, and agent
relationships form, change and decay
♦ When it is important that agents form
organizations and when adaptation and learning
are important at the organization level
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
22
Applicability (2/2)
♦ When it is important that agents have a spatial
component to their behaviors and interactions
♦ When the past is no predictor of the future
because the processes of growth and changes are
dynamic
♦ When scale-up to arbitrary levels is important
♦ When process structural change needs to be an
endogenous result of the model, rather than an
input to the model
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
23
Agent Simulation Vs Macro Simulation (1/3)
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
24
Agent Simulation Vs Macro Simulation (2/3)
 Can deal with multi-agent
 Differential equations are
systems directly because real
a well understood,
agent are represented by
established mathematical
simulated agent
framework
 Facilitates structural validation
 Elegant treatment of variable
structures
 Easy to document
 Allows to model adaptation
and evolution
 Easy to model heterogeneous
space and population
 Low number of
parameters, global input Provides different levels of
output behavior
observation
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
25
Agent Simulation Vs Macro Simulation (3/3)
 Development of complex
 Assumes homogeneous
models can be very costly
space and population
 Difficult to determine
minimal model
 Established formalism is
 No representation of the
missing, difficult to
individual and its locality, i.e.,
document
no conditional behavior, no
 Calibration problem, i.e., is
adaptive behavior, no flexible
difficult to find the best
interaction
parameter setting for a
model (given a structurally
valid model)
 Sensitivity problem, i.e.,
 Can only observe the system
even small changes may
as a whole, not its parts
have a large effect
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
26
Applications
Business and Organization
Society and Culture
Manufacturing operations
Ancient civilizations
Supply chains
Civil disobedience
Consumer markets
Social determinant of terrorism
Insurance industry
Organizational networks
Economics
Military
Artificial financial markets
Command and control
Trade networks
Force on force
Infrastructure
Biology
Electric power markets
Population dynamics
Transportation
Ecological networks
Hydrogen infrastructure
Animal group behavior
Crowds
Cell behavior and sub-cell processes
Pedestrian movement
Evacuation modelling
Prof. Agostino Poggi
Entertainment
Movies and games
Agent-Based Modelling
University of Parma
-
27
Software (1/2)
AgentSheets
AndroMeta
AnyLogic
Ascape
Breve
Cormas
DEVS
EcoLab
FLAME
JAS
Prof. Agostino Poggi
LSD
MAML
MATSim
MASON
MASS
MetaABM
MIMOSE
MobiDyc
Modelling4all
NetLogo
Agent-Based Modelling
RePast
Repast Simphony
SimPack
SimPy
SOARS
StarLogo
SugarScape
Swarm
VisualBots
Xholon
University of Parma
-
28
Software (2/2)
Prof. Agostino Poggi
Agent-Based Modelling
University of Parma
-
29
Download