Introduction Lecture 01
-
Upload
christian-andrew -
Category
Documents
-
view
219 -
download
0
Transcript of Introduction Lecture 01
-
7/29/2019 Introduction Lecture 01
1/47
LECTURE 1: INTRODUCTIONLECTURE 1: INTRODUCTION
Multiagent Systems
Based on An Introduction to MultiAgent
Systems by Michael Wooldridge, John Wiley &Sons, 2002.http://www.csc.liv.ac.uk/mjw/pubs/imas/
1
-
7/29/2019 Introduction Lecture 01
2/47
OverviewOverview
Five ongoing trends have marked the history
of computing:ubiquity;
intelligence;
delegation; and
human-orientation
2
-
7/29/2019 Introduction Lecture 01
3/47
UbiquityUbiquity
The continual reduction in cost of computing
capability has made it possible to introduceprocessing power into places and devices that would
have once been uneconomic
What could benefit from having a processor
embedded in it?
3
-
7/29/2019 Introduction Lecture 01
4/47
InterconnectionInterconnection Computer systems today no longer stand alone, but are
networked into large distributed systems
The internet is an obvious example, but networking is
spreading its ever-growing tentacles
Distributed and concurrent systems have become the
norm.
4
-
7/29/2019 Introduction Lecture 01
5/47
IntelligenceIntelligence The complexity of tasks that we are capable of
automating and delegating to computers hasgrown steadily
If you dont feel comfortable with this
definition of intelligence, its probably
because you are a human
5
-
7/29/2019 Introduction Lecture 01
6/47
DelegationDelegation Computers are doing more for us without our
intervention We are giving controlto computers, even in safety
critical tasks
One example: fly-by-wire aircraft, where the machinesjudgment may be trusted more than an experienced
pilot
Next on the agenda: fly-by-wire cars, intelligent brakingsystems, cruise control that maintains distance from
car in front
6
-
7/29/2019 Introduction Lecture 01
7/47
Human OrientationHuman Orientation The movement away from machine-oriented views
of programming toward concepts and metaphorsthat more closely reflect the way we ourselves
understand the world
Programmers an users re ate to t e mac nedifferently
Programmers conceptualize and implement software
in terms of higher-level more human-oriented abstractions
7
-
7/29/2019 Introduction Lecture 01
8/47
Programming progressionProgramming progression Programming has progressed through:
machine code;
assembly language;
-
sub-routines;
procedures & functions;
abstract data types;
objects;
to agents.8
-
7/29/2019 Introduction Lecture 01
9/47
Global ComputingGlobal Computing What techniques might be needed to deal
with systems composed of 1010 processors? Dont be deterred by its seeming to be
Hundreds of millions of people connected by
email once seemed to be science fiction
Lets assume that current softwaredevelopment models cant handle this
9
-
7/29/2019 Introduction Lecture 01
10/47
Where does it bring us?Where does it bring us? Delegation and Intelligence imply the need to build
computer systems that can act effectively on ourbehalf
This implies: The ability of computer systems to act independently
The ability of computer systems to act in a way that
represents our best interests while interacting with otherhumans or systems
10
-
7/29/2019 Introduction Lecture 01
11/47
Interconnection and DistributionInterconnection and Distribution Interconnection and Distribution have become core
motifs in Computer Science But Interconnection and Distribution, coupled with
the need for systems to represent our best interests,
mp es systems t at can cooperate an reacagreements (or even compete) with other systems
that have different interests (much as we do with
other people)
11
-
7/29/2019 Introduction Lecture 01
12/47
So Computer Science expandsSo Computer Science expands These issues were not studied in Computer
Science until recently All of these trends have led to the emergence
multiagent systems
12
-
7/29/2019 Introduction Lecture 01
13/47
Agents, a DefinitionAgents, a Definition An agent is a computer system that is
capable ofindependentaction on behalf ofits user or owner (figuring out what needs to
be done to satisfy design objectives, rather
than constantly being told)
13
-
7/29/2019 Introduction Lecture 01
14/47
Multiagent Systems, a DefinitionMultiagent Systems, a Definition A multiagent system is one that consists of a
number of agents, which interactwith one-another
In the most general case, agents will be acting on
behalf of users with different goals andmotivations
To successfully interact, they will require the
ability to cooperate, coordinate, and negotiatewith each other, much as people do
14
-
7/29/2019 Introduction Lecture 01
15/47
Agent Design, Society DesignAgent Design, Society Design The course covers two key problems:
How do we build agents capable of independent, autonomous
action, so that they can successfully carry out tasks we delegate to
them?
How do we build agents that are capable of interacting (cooperating,coordinating, negotiating) with other agents in order to successfully
carry out those delegated tasks, especially when the other agents
cannot be assumed to share the same interests/goals?
agent design, and society design (micro/macro)
15
-
7/29/2019 Introduction Lecture 01
16/47
Multiagent SystemsMultiagent Systems In MAS, we address questions such as:
How can cooperation emerge in societies of self-interested agents?
What kinds of languages can agents use to communicate?
How can self-interested agents recognize conflict, andhow can they (nevertheless) reach agreement?
How can autonomous agents coordinate their activities so
as to cooperatively achieve goals?
16
-
7/29/2019 Introduction Lecture 01
17/47
Multiagent SystemsMultiagent Systems
While these questions are all addressed in
part by other disciplines (notablyeconomics and social sciences), what
emphasizes that the agents in question arecomputational, information processing
entities.
17
-
7/29/2019 Introduction Lecture 01
18/47
The Vision ThingThe Vision Thing Its easiest to understand the field of MAS if you understand
researchers vision of the future
Fortunately, different researchers have different visions
The amalgamation of these visions (and research directions,
and methodologies, and interests, and) define the field
But the fields researchers clearly have enough in common toconsider each others work relevant to their own
18
-
7/29/2019 Introduction Lecture 01
19/47
Spacecraft ControlSpacecraft Control When a space probe makes its long flight from Earth to the
outer planets, a ground crew is usually required to
continually track its progress, and decide how to deal with
unexpected eventualities. This is costly and, if decisions are
, .
reasons, organizations like NASA are seriously investigating
the possibility of making probes more autonomous
giving them richer decision making capabilities and
responsibilities. This is not fiction: NASAs DS1 has done it!
19
-
7/29/2019 Introduction Lecture 01
20/47
Deep Space 1Deep Space 1 http://nmp.jpl.nasa.gov/ds1/
Deep Space 1launched from CapeCanaveral on October 24,1998. During a highly
,
it tested 12 advanced, high-risk technologies in space. Inan extremely successful extended mission, it encounteredcomet Borrelly and returned the best images and otherscience data ever from a comet. During its fully successfulhyperextended mission, it conducted further technologytests. The spacecraft was retired on December 18, 2001.
NASA Web site
20
-
7/29/2019 Introduction Lecture 01
21/47
Autonomous Agents for specialized tasksAutonomous Agents for specialized tasks
The DS1 example is one of a generic class
Agents (and their physical instantiation in robots)
have a role to play in high-risk situations, unsuitable
or impossible for humans
The degree of autonomy will differ depending on
the situation.
21
-
7/29/2019 Introduction Lecture 01
22/47
Air Traffic ControlAir Traffic Control A key air-traffic control systemsuddenly fails, leaving
flights in the vicinity of the airport with no air-traffic
control support. Fortunately, autonomous air-traffic
control systems in nearby airports recognize the failure of
their peer, and cooperate to track and deal with all
affected flights.
Systems taking the initiative when necessary
Agents cooperating to solve problems beyond the
capabilities of any individual agent
22
-
7/29/2019 Introduction Lecture 01
23/47
Internet AgentsInternet Agents After the wettest and coldest UK winter on record,
you are in desperate need of a last minute holidaysomewhere warm and dry. After specifying your
requirements to your personal digital assistant
, converses w a num er o eren e
sites, which sell services such as flights, hotel rooms,
and hire cars. After hard negotiation on your behalf
with a range of sites, your PDA presents you with
package holiday.
23
-
7/29/2019 Introduction Lecture 01
24/47
Research IssuesResearch Issues How do you state your preferences to your agent?
How can your agent compare different deals from
different vendors? What if there are many different
parameters?
What algorithms can your agent use to negotiate
with other agents (to make sure you get a gooddeal)?
24
-
7/29/2019 Introduction Lecture 01
25/47
MAS is InterdisciplinaryMAS is Interdisciplinary The field of MAS is influenced and inspired by many
other fields:
Economics
osop y
Game Theory
Logic
Ecology
Social Sciences
This can be both a strength and a weakness.25
-
7/29/2019 Introduction Lecture 01
26/47
Some Views of the FieldSome Views of the Field Agents as a paradigm for software engineering:
Software engineers have derived a progressively better
understanding of the characteristics of complexity in software.
It is now widel reco nized that interaction is robabl the
most important single characteristic of complex software
Over the last two decades, a major Computer Science research
topic has been the development of tools and techniques to
model, understand, and implement systems in whichinteraction is the norm
26
-
7/29/2019 Introduction Lecture 01
27/47
Some Views of the FieldSome Views of the Field Agents as a tool for understanding human
societies:
Multiagent systems provide a novel new tool fors mu at ng soc et es, w c may e p s e somelight on various kinds of social processes.
This has analogies with the interest in theories of
the mind explored by some artificial intelligenceresearchers
27
-
7/29/2019 Introduction Lecture 01
28/47
Objections to MASObjections to MAS Isnt it all just Distributed/Concurrent
Systems?
Agents are assumed to be autonomous, capableo ma ng n epen ent ec s on so t ey neemechanisms to synchronize and coordinate theiractivities at run time
In a MAS each agent is (can be) self-interested,oppose to distributed/concurrent systems.
28
-
7/29/2019 Introduction Lecture 01
29/47
Objections to MASObjections to MAS Isnt it all just AI?
We dont need to solve all the problems ofartificial intelligence (i.e., all the components of
intelligence) in order to build really useful agents.
Intelligent agents are 99% computer science and1% IA Oren Etzioni.
Classical AI ignored socialaspects of agency. Theseare important parts of intelligent activity in real-
world settings.29
-
7/29/2019 Introduction Lecture 01
30/47
Objections to MASObjections to MAS Isnt it all just Economics/Game Theory?
These fields also have a lot to teach us in multiagentsystems, but:
Insofar as game theory provides descriptive
concepts, it doesnt always tell us howtocompute solutions; were concerned with
computational, resource-bounded agents
Some assumptions in economics/game theory(such as a rational agent) may not be valid or
useful in building artificial agents
30
-
7/29/2019 Introduction Lecture 01
31/47
Objections to MASObjections to MAS Isnt it all just Social Science?
We can draw insights from the study of humansocieties, but there is no particular reason to
believe that artificial societies will be constructed
in the same way
Again, we have inspiration and cross-fertilization,
but hardly subsumption
31
-
7/29/2019 Introduction Lecture 01
32/47
What is an Agent?What is an Agent? The main point about agents is they are autonomous:
capable of acting independently, exhibiting control over
their internal state
Thus: an agent is a computer system capable of
autonomous action in some environment in order to meet
its design objectives
32
SYSTEM
ENVIRONMENT
input output
-
7/29/2019 Introduction Lecture 01
33/47
What is an Agent?What is an Agent?
SYSTEM
ENVIRONMENT
input output
The key problem facing an agent is that of deciding which
of its actions it should perform in order to best satisfy its
design objectives.
33
-
7/29/2019 Introduction Lecture 01
34/47
What is an Agent?What is an Agent? Trivial (non-interesting) agents:
thermostat
UNIX daemon (e.g., biff)
An intelligent agent is a computer system capable of
ex e au onomous ac on n some env ronmen
Byflexible, we mean:
reactive
pro-active
social
34
-
7/29/2019 Introduction Lecture 01
35/47
ReactivityReactivity If a programs environment is guaranteed to be fixed, the program need
never worry about its own success or failure program just executes
blindly Example of fixed environment: compiler
, .
Many (most?) interesting environments are dynamic
Software is hard to build for dynamic domains: program must take into
account possibility of failure ask itself whether it is worth executing!
A reactive system is one that maintains an ongoing interaction with its
environment, and responds to changes that occur in it (in time for the
response to be useful)35
-
7/29/2019 Introduction Lecture 01
36/47
ProactivenessProactiveness Reacting to an environment is easy (e.g., stimulus
response rules) But we generally want agents to do things for us
Hence goal directed behavior
Pro-activeness = generating and attempting toachieve goals; not driven solely by events; taking
the initiative
Recognizing opportunities
36
-
7/29/2019 Introduction Lecture 01
37/47
Balancing Reactive and GoalBalancing Reactive and Goal--OrientedOriented
BehaviorBehavior
We want our agents to be reactive, responding to
changing conditions in an appropriate (timely) fashion
We want our agents to systematically work towards
long-term goals
These two considerations can be at odds with oneanother
Designing an agent that can balance the two remains
an open research problem
37
-
7/29/2019 Introduction Lecture 01
38/47
Social AbilitySocial Ability The real world is a multi-agent environment: we cannot go
around attempting to achieve goals without taking others
into account
Some goals can only be achieved with the cooperation of
others
Similarly for many computer environments: witness theInternet
Social abilityin agents is the ability to interact with other
agents (and possibly humans) via some kind ofagent-
communication language, and perhaps cooperate with
others
38
-
7/29/2019 Introduction Lecture 01
39/47
Other PropertiesOther Properties Other properties, sometimes discussed in the context of
agency:
mobility: the ability of an agent to move around an electronic
network
veracity: an agent will not knowingly communicate false
information
benevolence: agents do not have conflicting goals, and that
every agent will therefore always try to do what is asked of it
rationality: agent will act in order to achieve its goals, and will
not act in such a way as to prevent its goals being achieved at
least insofar as its beliefs permit
learning/adaption: agents improve performance over time
39
-
7/29/2019 Introduction Lecture 01
40/47
EnvironmentsEnvironments AccessibleAccessible vs.vs. inaccessibleinaccessible
An accessible environment is one in which the
agent can obtain complete, accurate, up-to-dateinformation about the environments state
The more accessible an environment is, the simplerit is to build agents to operate in it
40
-
7/29/2019 Introduction Lecture 01
41/47
EnvironmentsEnvironments
DeterministicDeterministic vs.vs. nonnon--deterministicdeterministic
A deterministic environment is one in which any
action has a single guaranteed effect there is no
uncertainty about the state that will result from
performing an action
Non-deterministic environments present greater
problems for the agent designer
41
-
7/29/2019 Introduction Lecture 01
42/47
EnvironmentsEnvironments -- StaticStatic vs.vs. dynamicdynamic A static environment is one that can be assumed to remain
unchanged except by the performance of actions by the agent
A dynamic environment is one that has other processes
operating on it, and which hence changes in ways beyond the
agents control
Other processes can interfere with the agents actions (as in
concurrent systems theory)
The physical world is a highly dynamic environment
42
-
7/29/2019 Introduction Lecture 01
43/47
Agents and ObjectsAgents and Objects Are agents just objects by another name?
Object:encapsulates some state
has methods, corresponding to operations
that may be performed on this state
43
-
7/29/2019 Introduction Lecture 01
44/47
Agents and ObjectsAgents and Objects Main differences:
agents are autonomous:
agents embody stronger notion of autonomy than objects,
and in particular, they decide for themselves whether or not
to perform an action on request from another agent
agents are smart:
capable of flexible (reactive, pro-active, social) behavior, and
the standard object model has nothing to say about such
types of behavior
agents are active:a multi-agent system is inherently multi-threaded, in that
each agent is assumed to have at least one thread of active
control44
-
7/29/2019 Introduction Lecture 01
45/47
Objects do it for freeObjects do it for free agents do it because they want to
agents do it for money
45
-
7/29/2019 Introduction Lecture 01
46/47
Agents and Expert SystemsAgents and Expert Systems Arent agents just expert systems by another name?
Expert systems typically disembodied expertise about
some (abstract) domain of discourse (e.g., blood
diseases)
xamp e: nows a out oo seases n
humans
It has a wealth of knowledge about blood diseases, in the
form of rules
A doctor can obtain expert advice about blood diseases bygiving MYCIN facts, answering questions, and posing queries
46
-
7/29/2019 Introduction Lecture 01
47/47
Agents and Expert SystemsAgents and Expert Systems Main differences:
agents situated in an environment:MYCIN is not aware of the world only
information obtained is by asking the user
questionsagents act:
MYCIN does not operate on patients
Some real-time (typically process control)expert systems are agents
47