lect01 introtoagents10 100201051415 phpapp01
TRANSCRIPT
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
1/59
LECTURE 1:
INTRODUCTION TO
AGENTS AND MAS
Artificial Intelligence II Multi-Agent Systems
Introduction to Multi-Agent Systems
URV, Winter-Spring 2010
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
2/59
Outline of the lecture
Main trends in Computer Science
Agents and multi-agent systems Viewpoints on agent technology
Agent technology roadmap Challenges on agent technology
Objections to MAS
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
3/59
Important aspects in Computer Science
Five ongoing trends are marking the history
of computing:
ubiquity
interconnection
intelligence delegation
human-orientation in programming methodologies
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
4/59
Ubiquity
The continuous reduction in cost of
computing capability has made it possible to
introduce processing power into places anddevices that would have once been
uneconomic / unimaginable [e.g. Nike+]
As processing capability spreads,
sophistication (and intelligence of a sort)
becomes ubiquitous What could benefit from having a processor
embedded in it?
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
5/59
Ambient Intelligence
Aimed at seamless delivery of services and applications.
Relies on ubiquitous computing, ubiquitouscommunication and intelligent user interfaces.
The vision An environment of potentially thousands of embedded and
mobile devices (or software components) interacting to supportuser-centred goals and activity.
Suggests a component-oriented view of the world in which thecomponents are independent and distributed.
Autonomy, distribution, adaptation, responsiveness, and so on,are key characteristics of these components, and in this sense
they share the same characteristics as agents.
Requires agents to be able to interact with numerousother agents in the environment around them in order to
achieve their goals.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
6/59
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
7/59
Interconnection
Computer systems today no longer stand
alone, but are networked into large
distributed systems Obvious example: Internet
Since distributed and concurrent systemshave become the norm, some researchers
are putting forward theoretical models that
portray computing as primarily a process ofinteraction
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
8/59
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
9/59
Connectivity -Technological context
Semantic Web
The Semantic Web is based on the idea that the data
on the Web can be defined and linked in such a waythat it can be used by machines for the automatic
processing and integration of data across different
applications (Berners-Lee et al., 2001).
Peer-to-Peer computing
Grid computing
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
10/59
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
11/59
Intelligence
The complexity of tasks that we are capable
of automating and delegating to computers
has grown steadily
If you dont feel comfortable with this
definition of intelligence, its probablybecause you are a human
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
12/59
Delegation
Computers are doing more for us withoutour intervention
We are giving control to computers, even insafety critical tasks
One example: fly-by-wire aircraft, where the
machines judgment may be trusted more
than an experienced pilot
Next on the agenda: fly-by-wire cars,intelligent braking systems, cruise control that
maintains distance from car in front
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
13/59
Driving automation
1. Road ConditionReporting2. Adaptive Cruise
Control3. OmnidirectionalCollision System
4. Lane-DeparturePrevention5. Auto Parallel Park
6. Blind-SpotSensors7. Corner Speed
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
14/59
Human Orientation
The movement away from machine-oriented
views of programming toward concepts and
metaphors that more closely reflect the way
we ourselves understand the world
Programmers (and users!) relate to themachine differently
Programmers conceptualize and implement
software in terms of higher-level more
human-oriented abstractions
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
15/59
Agent-Oriented Software Engineering
Programming has progressed through
machine code
assembly language
machine-independent programming languages
sub-routines procedures & functions
abstract data types
objects
to agents.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
16/59
Where does it bring us?
Delegation and Intelligence imply the need to
build computer systems that can act
effectively on our behalf
This implies:
The ability of computer systems to actindependently
The ability of computer systems to act in a way
that represents our best interests while interactingwith other humans or systems
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
17/59
Interconnection 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, implies systems that cancooperate and reach agreements (or even
compete) with other systems that have
different interests (much as we do with otherpeople)
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
18/59
So Computer Science expands
These issues were not studied in Computer
Science until recently All of these trends have led to the emergence
of a new field in Computer Science:
multiagent systems
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
19/59
Agents, a Definition
An agent is a computer system that is located in adynamic environment and is capable ofindependent action on behalf of its user or owner(figuring out what needs to be done to satisfydesign objectives, rather than constantly beingtold)
Intelligent action, probably using ArtificialIntelligence tools and techniques
An agent should satisfy a certain number of
properties [lecture on week 3]
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
20/59
Multiagent
Systems, a Definition A multiagent system is one that
consists of a number of agents,which interact with one-another
In the most simple case, allagents are programmed by thesame team and they collaborateto complete a task
In the most general case,agents will be acting on behalfof users with different goals andmotivations
To successfully interact, theywill require the ability tocooperate, coordinate, andnegotiate with each other, much
as people do
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
21/59
Related areas
Several sub-disciplines of informationtechnology are related to software agenttechnology:
computer networks,
software engineering,
Artificial Intelligence, human-computer interaction,
distributed and concurrent systems,
mobile systems, computer-supportedcooperative work, control systems, decisionsupport, information retrieval and management,electronic commerce...
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
22/59
Micro and macro aspects
Agent technologies can be grouped into threecategories Agent-level [micro level]
technologies and techniques concerned only with individualagents for example, procedures for agent reasoning andlearning.
Interaction-level
technologies and techniques that concern thecommunications between agents
communication languages, interaction protocols andresource allocation mechanisms.
Organization-level [macro level]
technologies and techniques related to agent societies as awhole.
structure, trust, norms, obligations, etc.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
23/59
Multiagent
Systems
In Multiagent Systems, we address questions
such as:
What kinds of languages can agents use tocommunicate?
How can cooperation emerge in societies of self-
interested agents? How can self-interested agents recognize conflict,
and how can they (nevertheless) reach
agreement?
How can autonomous agents coordinate their
activities so as to cooperatively achieve goals?
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
24/59
Multiagent
Systems
While these questions are all addressed
in part by other disciplines (notably
Economics and Social Sciences), whatmakes the multiagent systems field
unique is that it emphasizes that theagents in question are computational,
information processing entities.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
25/59
Example 1 -
Spacecraft
Control
When a space probe makes a flight 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 required quickly, it is simply not
practicable. NASA is investigating the possibility ofmaking probes more autonomous giving them
richer decision making capabilities and
responsibilities. This is not fiction: NASAs DS1 has done it!
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
26/59
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
27/59
Autonomous Agents for specialized tasks
The DS1 example is one of a generic class
Agents (and their physical instantiation inrobots) have a role to play in high-risk
situations, unsuitable or impossible for
humans The degree of autonomy will differ depending
on the situation (remote human control maybe an alternative, but not always)
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
28/59
Example 2 -
Air
Traffic Control
A key air-traffic control system suddenly 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 allaffected flights.
Systems taking the initiative when necessary Agents cooperating to solve problems beyond
the capabilities of any individual agent
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
29/59
Example 3 -
Internet
Agents
Searching the Internet for the answer to a
specific query can be a long and tedious
process. So, why not allow a computerprogram an agent do searches for us?
The agent would typically be given a query
that would require synthesizing pieces of
information from various different Internet
information sources.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
30/59
Example 4
e-personal assistants
Internet agents need not simply search
They can plan, arrange, buy, negotiate
carry out arrangements of all sorts that wouldnormally be done by their human user
As more can be done electronically, softwareagents theoretically have more access to
systems that affect the real-world
For example, preparing a holiday trip
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
31/59
Research Issues
How do you state your preferences to youragent?
How can your agent compare different dealsfrom different vendors? What if there are manydifferent parameters?
What algorithms can your agent use tonegotiate with other agents?
These issues arent frivolous automatedprocurement could be used massively by (forexample) government agencies
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
32/59
Multiagent
Systems is Interdisciplinary
The field of Multiagent Systems is influenced andinspired by many other fields:
Economics
Philosophy
Game Theory
Logic
Ecology
Social Sciences
This can be both a strength (infusing well-founded
methodologies into the field) and a weakness (there
are many different views as to what the field is about)
This has analogies with Artificial Intelligence itself
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
33/59
Some Views of the Field
Agents as a paradigm for software
engineering:
Software engineers have derived aprogressively better understanding of the
characteristics of complexity in software. It
is now widely recognized that interaction isprobably the most important single
characteristic of complex software
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
34/59
Agents as design metaphor
Agent-oriented software engineering
Agents allow software designers and developersto structure an application using autonomous,communicative components.
In this sense, software agents offer a new and
often more appropriate route to the developmentof complex computational systems, especially inopen and dynamic environments.
Modularity, reusability,collaborative behaviour
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
35/59
Agents as source of technologies
Software agent technologies span a range ofspecific techniques and algorithms
balancing reaction and deliberation in individualagent architectures [lecture 2]
learning from and about other agents in the
environment eliciting and acting upon user preferences
finding ways to negotiate and cooperate with
other agents developing appropriate means of forming and
managing coalitions (and other organizations)
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
36/59
Social simulation
Agents as a tool for understanding human
societies:
Multiagent systems provide a novel newtool for simulating societies, which may
help shed some light on various kinds of
social processes.
This has analogies with the interest in
theories of the mind explored by someArtificial Intelligence researchers
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
37/59
Agent-based simulation
Multi-agent systems provide strong models
for representing complex and dynamic real-
world environments. modelling of the impact of climate change on
biological populations
modelling the impact of public policy options onsocial or economic behaviour
modelling intelligent buildings
modelling traffic systems
modelling biological populations
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
38/59
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
39/59
Distributed computing theory
Multiagent Systems is primarily a search for
appropriate theoretical foundations:
We want to build systems of interacting,autonomous agents, but we dont yet know
what these systems should look like
You can take a neat or scruffy approach tothe problem, seeing it as a problem of theory
or a problem of engineering This, too, has analogies with Artificial
Intelligence research
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
40/59
Agent technology roadmap [AgentLink]
There are four broad phases Phase 1: current
Phase 2: short-term future
Phase 3: medium-term future
Phase 4: long-term future
Time phases distinguished along dimensions: The degree to which the participating agents share
common domain knowledge and common goals.
The degree to which participating agents aredesigned by the same or diverse design teams.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
41/59
Phase 1: present state
Systems are designed by the same teams.
Systems share common domain knowledge.
Participating agents have a global objective.
Agent communication languages are based onstandards.
Interaction protocols remain non-standard.
Low scalability, few agents.
Example systems include those to enable
automated scheduling coordination betweendifferent departments, etc.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
42/59
Phase 2: short-term future
Multi-agent systems will increasingly be designed to crosscorporate boundaries.
Participating agents have few goals in common, althoughtheir interactions will still concern a common domain
Agents will be designed by a same team, and will sharecommon domain knowledge.
Standard agent communication languages, such as FIPA
ACL, will be used. Interaction protocols will be mixed between standard and
non-standard ones.
Systems will be able to handle large numbers of agents inpre-determined environments.
Development methodologies, languages and tools will havereached a degree of maturity.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
43/59
Phase 3: medium-term future
Multi-agent systems will permit participation byheterogeneous agents, designed by different designers orteams.
Any agent will be able to participate in these systems.
Their (observable) behaviour conforms to publicly-stated requirements and standards.
Open systems will typically be specific to particular
application domains, such as B2B eCommerce orbioinformatics.
The languages and protocols used will be agreed andstandardised, perhaps drawn from public libraries of
alternative protocols. Commonly agreed modelling languages
Agent-UML
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
44/59
Phase 4: Long-term future
Truly open and fully-scalable agent systemsspanning multiple application domains and
developed by diverse design terms Agents learn appropriate protocols and behaviour
upon entry to system
Languages, protocols, and behaviours emergefrom actual agent interactions.
Evolving organisational structure with multiple,dynamic, interacting organisations.
Self-modifying agent communications languages.
Use of agent-specific design methodologies.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
45/59
Challenges in agent technology
Implementing multi-agent systems is still a complextask. Lack of maturity in both methodologies and programming
tools.
Lack of specialized debugging tools.
Lack of skills needed to move from analysis and design tocode.
Problems associated with awareness of the specifics ofdifferent agent platforms.
Problems in understanding the nature of what is a new anddistinct approach to systems development.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
46/59
Technological Challenges
Increase quality of agent software to industrialstandard
Agent oriented design methodologies, tools anddevelopment environments.
Seamless integration with existing technologies.
Web, information systems
Non-functional properties: scalability, performance,reliability, and robustness.
Analysis point of view
Agent technology requires dedicated basicconcepts and languages: Dynamic aspects, such as time and action.
Locality aspects, such as position in a space.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
47/59
Technological Challenges
Increase quality of agent software to industrial
standard (cont.)
Design point of view The key point is reusability
Libraries of
Generic organisation models (hierarchical, flat, etc.).
Generic agent models (reactive, deliberative, etc.)
Generic task models (diagnostic, information filtering, etc.).
Communication languages and patterns for agent societes.
Ontology patterns for agent requirements, models, and organisation.
Interaction protocol patterns between agents with special roles. Reusable knowledge bases.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
48/59
Technological Challenges
Develop reasoning capabilities for agents
Virtual organisations, e-institutions
Heterogeneity of agents, trust, and failurehandling and recovery.
Norms, legislation, authorities, and enforcement.
Coalition formation Dynamic coalitions of small groups of agents.
Negotiation and argumentation strategies
Distributed planning techniques
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
49/59
Technological Challenges
Develop agent ability to understand userrequirements
User profiling Personalisation
Knowledge acquisition tools
Develop agent ability to adapt to changesin environment
Learning
Evolutionary programming techniques
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
50/59
Technological Challenges
Ensure user confidence and trust in agents
Security technologies. Security in open agent systems.
Agents must be secure and must not reveal informationinappropriately.
Confidence that agents representing other parties willbehave within certain constrains.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
51/59
Technological Challenges
Ensure user confidence and trust in agents(cont).
Models and infrastructure for trust and reputation. Reputation mechanisms to assess the past behaviour of
particular agents or users, to allow avoidance ofuntrustworthy agents.
The adoption of norms (social rules) by all members ofan open agent system.
Enforcement of sanctions against agents not followingrules (or self-enforcing protocols).
The use of electronic contracts to represent and enforceagreements between several parties.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
52/59
Technological Challenges Provide semantic infrastructure for open agentcommunities
Ontologies
Semantic Web
Matchmaking and broker architectures
Exiting out of laboratories requires A greater understanding of how agents, databases and
information systems interact. Investigation of the real-world implications of information
agents.
Development of benchmarks for agent system performance
and efficiency. New web standards that enable structural and semantic
description of information.
Creation of common ontologies, thesauri, or knowledge
bases.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
53/59
Preconditions for wide-scale usage
Software agent technology is better than other ones.
At least, when compared using one measurable
parameter in the application field ....
There are enough trained, skillful persons to
develop applications with software agent
technology.
Development tools are easy to use and powerful
enough to develop applications.
Technology is generally known.
Including top-level management.
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
54/59
Objections to MAS (I)
Isnt it all just Distributed/Concurrent Systems?
There is much to learn from this community,
but: Agents are assumed to be autonomous, capable of
making independent decision so they need
mechanisms to synchronize and coordinate theiractivities at run time
Agents are (can be) self-interested, so their
interactions are economic encounters(negotiation)
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
55/59
Objections to MAS (II)
Isnt it all just AI?
We dont need to solve all the problems of
Artificial Intelligence (i.e., all the components ofintelligence) in order to build really useful agents
Classical AI ignored social aspects of agency.
These are important parts of intelligent activity inreal-world settings
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
56/59
Objections to MAS (III)
Isnt it all just Economics/Game Theory?These fields also have a lot to teach us in
multiagent systems, but:
Insofar as game theory provides descriptive
concepts, it doesnt always tell us how to compute
solutions; were concerned with computational,resource-bounded agents
Some assumptions in economics/game theory
(such as a rational agent) may not be valid oruseful in building artificial agents
b A
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
57/59
Objections to MAS (IV)
Isnt it all just Social Science?
We can draw insights from the study of
human societies, but there is no particularreason to believe that artificial societies will be
constructed in the same way
Again, we have inspiration and cross-fertilization, but hardly subsumption
R di f hi k
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
58/59
Readings for this week
M.Wooldridge:An
introduction to MultiAgent
Systems chapter 1 A.Mas:Agentes software y
sistemas multi-agente:
conceptos, arquitecturas yaplicaciones initial part of
chapter 1
Agent Technology:
Computing as Interaction
AgentLink III roadmap
Mi d h 1 (W ld id )
-
7/26/2019 Lect01 Introtoagents10 100201051415 Phpapp01
59/59
Mindmap, chapter 1 (Wooldridge)