SlideShare a Scribd company logo
Chapter 6
Agent Communication
Marc-Philippe Huget
Introduction
 Agent communication is essential for facilitating cooperation,
coordination, and other aspects of agent interaction
 Two major approaches to agent communication
 Direct communication, and
 Indirect communication
 Early work on agent communication focused on direct
communication with the use of different protocols
 Protocols are efficient means for information exchange and
coordination; however, they restrict agents’ autonomy; agents
have to strictly follow the admissible sequence of messages
imposed by the protocol.
Introduction
 Agent communication is based on human conversations.
 Message contains a verb and a content.
 For instance: inform (it rains) inform is the verb and it rains
is the content. F(P): F- force; P- message content
 Several types of speech acts exist in human conversations:
 Representatives, which give information
 Directives, which require something from the recipient
 Commissives, which commit the sender to do something in the
future
 Expressives, which describe the mental states of the sender
 Declarations, which process an act just by saying it
 Agents do not use all the types of verbs present in a natural
language but a smaller subset.
Direct Communication
 Communication between cognitive agents (agents that have a
representation of themselves, of other agents, and of the
world, and reason about their courses of action) where the
recipient is formally defined
 Several elements require attention when considering direct
communication:
 an agent communication language
 ontologies
 communication support
Indirect Communication
 Message is released in the environment. Any agent sufficiently
close to it can perceive it.
 Indirect communication is used by reactive agents (agents that
do not have a representation of themselves, of other agents,
and of the world.
 Can only be used for limited coordination, since it is not
possible to represent a wide range of messages.
 Examples of indirect communication include pheromones
(example the MANTA project ) and forces (AI, robotics, and
multi-agent systems).
Agent Communication Versus
Multi-agent Communication
 Agent communication (as opposed to multi-agent
communication) is a one-to-one communication in which a
speaker utters speech acts to a single addressee.
 Several features characterize one-to-one communication:
 Communication is between two agents: a speaker and an
addressee.
 Communication is privileged between two agents, that is, it is
not possible for other agents to hear the communication.
 Entering and leaving communications are subject to agreement
from other
 Terms of termination of communication are necessarily known
before beginning the communication.
Agent Communication Versus
Multi-agent Communication
 Multi-agent communication allow for indirect communication.
That is, it should allow agents to listen to messages transmitted
in their vicinity (either physical or logical) but not too far-away
communication.
 It is then important to add the notion of earshot. In human
communication, earshot can range from whispering (private
communications) to shouting where even further agents can
hear.
 Multi-agent systems are open systems; as a consequence,
agents can enter and leave at any moment during the
conversation.
 Termination of communications is not necessarily decided in
advance. Termination can happen if no more communication is
needed.
Direct Communication:
Agent Communication Languages
 Communicating requires the use of common language. For
example, we communicate using English, to understand each
other.
 Agent communication languages do not offer a rich set of
communicative acts as natural languages.
 Currently, there exist two main agent communication
languages: KQML (Knowledge Query and Manipulation
Language) and FIPA ACL.
 KQML allow data exchange between distributed knowledge
bases. KQML is composed of
 7 performatives for conversation management (sorry, error, etc.),
 17 performatives for the conversation (ask-if, tell, etc.), and
 11 performatives for conversation sending (forward, broker-all,
etc.).
KQML Performatives (abstracted from EIWG, 1994)
Direct Communication:
Agent Communication Languages
 KQML is the older communication languages and is gradually
replaced by FIPA ACL for several reasons:
 There was a need for clear semantics for the language, (make a
clear different between KQML and FIPA ACL)
 It was recommended to have a smaller set of communicative acts,
while keeping the autonomy of agents,
 FIPA ACL has implementation of various protocols, and
 FIPA ACL supports communication at various levels.
The KQML Language
 Knowledge Query and Manipulation Language (KQML) is a
language that is designed to support interaction among
intelligent software agents.
 It was developed by the ARPA-supported Knowledge Sharing
Effort and independently implemented by several research
groups.
 It has been successfully used to implement a variety of
information systems using different software architectures.
 The KQML language simplifies its implementation by allowing
KQML messages to carry arbitrary useful information, such as
the names and addresses of the sending and receiving agents,
a unique message identifier, and notations by any intervening
agents.
KQML: the 3 layers
The KQML language can be viewed as consisting of three layers: the content,
message and communication layers, as shown in Figure.
• The content layer specifies the actual content of the message, and the KQML
standard itself has nothing to say about this.
• The set of performatives provided by the language, and shown in Table,
constitute the message layer, which in turn forms the core of the language.
• This layer of abstraction provides the performative and specifies the protocol for
delivering the message that subsumes the content.
• KQML specifies several protocols, including synchronous (where a blocking
query waits for an expected reply) and asynchronous (which involves non-
blocking messages).
• The communication layer encodes low level communication parameters, such
as the identities of the sender and the recipient, and unique identifiers for the
particular speech act.
Example of an instantiated performative
A KQML message from agent joe representing a query about the
price of a share of IBM stock might be encoded as:
In this message, the KQML performative is ask-one, the
content is (price ibm ?price), the ontology assumed
by the query is identified by the token nyse-ticks, the
receiver of the message is to be a server identified as
stock-server and the query is written in a language called
LPROLOG. The value of the :content keyword is the content
level, the values of the :reply-with, :sender, :receiver
keywords form the communication layer
and the performative name, with the :language and
:ontology form the message layer.
Applications of KQML Language
 KQML has been used as the communication language in several
technology integration experiments in the ARPA Rome Lab Planning
Initiative.
 One of these experiments supported an integrated planning and
scheduling system for military transportation logistics linking a
planning agent, with a scheduler (in Common Lisp), a knowledge
base, and a case based reasoning tool (in Common Lisp).
 All of the components integrated were pre-existing systems which
were not designed to work in a cooperative distributed environment.
FIPA ACL
 The Foundation for Intelligent Physical Agents (FIPA) was formed in
1996 to produce software standards for heterogeneous and
interacting agents and agent-based systems.
 FIPA’s agent communication language, like KQML, is based on speech
act theory: messages are actions or communicative acts, as they are
intended to perform some action by virtue of being sent.
 FIPA ACL is superficially similar to KQML. Its syntax is identical to
KQML’s except for different names for some reserved primitives.
 Thus, it maintains the KQML approach of separating the outer
language from the inner language. The outer language defines the
intended meaning of the message; the inner, or content, language
denotes the variety of interesting agent architectures.
 Thus, KQML introduces a small number of KQML performatives that
agents use to describe the metadata specifying the information
requirements and capabilities; it also introduces a special class of
agents called communication facilitators.
 A facilitator is an agent that performs various useful communication
services, such as maintaining a registry of service names, forwarding
messages to named services, routing messages based on content,
matchmaking between information providers and clients, and
providing mediation and translation services.
Agent Communication
 This chapter will describe FIPA ACL.
 F(P); where F is force P is message content
 For example: inform (it rains); Force (F) is the performative verb
 Performatives in FIPA ACL are arranged into five categories:
 Passing information,
 Requesting information,
 Negotiation,
 Performing actions, and
 Error handling.
 However, we note that agent-based applications can be (and
have been) developed using traditional third generation
languages like Lisp, C, or Prolog, and object-oriented (OO)
languages such as C++ or Smalltalk.
Direct Communication:
Agent Communication Languages
Primitive
performatives
FIPA ACL
 Inform: an agent informs recipients about something it
believes and it believes recipients do not yet believe it.
 Request: request that recipient perform actions.
 KQML is lack of semantics associated with performative. FIPA
ACL addresses this issue via two parts:
 Feasible pre condition: pre condition that have to be evaluated to
true to execute the performative.
 Rational effect: expresses what the sender hopes to bring about.
 Semantics are expresses via FIPA SL (refer table 6.2)
Direct Communication:
Ontologies
 It is possible that two agents speak about the same concept
but not with the same term: using the terms “painter” and
“artist,” for instance.
 In order to avoid these differences, it is important to use a
common vocabulary also called an ontology
 An ontology defines the basic terms and relations comprising
the vocabulary of a topic area as well as the rules for
combining terms and relations to define extensions to the
vocabulary
 Example of ontology for computer science department is
shown next:
Direct Communication:
Ontologies
Direct Communication:
Communication Support
 Communicating requires a communication support, which is a
mechanism that stores, retrieves, and directs messages to the
agents.
 Communication support is present in agent platforms such as
JADE and MadKit , where it is possible to send messages to
 Agents,
 The roles they play in a group, or
 A group.
Direct Communication:
Structuring Communication
Three main different approaches exist to structure
direct communication:
1. Protocols: Main approach to represent communication in MAS
systems. The way to structure communication in a rigid manner.
A finite set of performatives is allowed for a particular state of
the interaction.
2. Dialogue games: offer more flexibility in communication in
comparison with protocols. The communicative acts are paired,
for instance, a question is paired with answer. It is possible to
embed dialogue games into other dialogue games.
3. Argumentation systems: agents challenge the content of each
message and as a consequence, agents have to justify their
message content via arguments. Argumentation is frequently
used in negotiation-based systems.
Protocol
 Research on protocols in multi-agent systems share work
with communication protocol engineering and interaction
protocol engineering.
 Interaction protocol engineering composed of several
phases from analysis to execution
Interaction Protocol Engineering
1. Analysis: This phase describes informally the protocol. The
exact sequence of messages, the content, and the different
requirements the protocol should follow.
2. Formal Description : To avoid misunderstandings and to
ease validation. Several formal description techniques are
available for interaction protocol designers:
 Finite State Machines : Based on graph theory, where each state
corresponds to an interaction state and a transition sends or
receives a message. Example include AgenTalk , and COOL
 Petri nets: Allow multiple flows and synchronization on the
same net.
 UML-based notations: Example FIPA UAML and Agent UML
 Others: temporal logic and Z language
Interaction Protocol Engineering
3. Validation: The validation can be structural (checking for
deadlock, termination, etc.) or behavioral (checking whether
the protocol achieves its goals).
Two techniques exist for the validation of interaction
protocols:
 Reachability analysis : A graph of all the reachable states of the
protocol is built-validation is based on graph properties.
 Model checking: A graph of reachable states is built, but this
time the property is expressed as a temporal logic formula.
 Another kind of validation is the agent communication
language compliance verification, which checks whether an
agent correctly implements the protocol and agent
communication language.
Interaction Protocol Engineering
4. Protocol Synthesis: The actual implementation. First the
skeleton of the protocol is generated. Then the semantics
associated with each transition is added. The first phase is
automatic and the second one is manual. System unable to
generate the semantics. The advantage is that it reduces the
effort required for generating a protocol since it s not necessary
to rewrite each time the semantics of the protocol modified.
5. Conformance Testing: Checks whether the implemented
protocol contains the requirements defined in the analysis
phase.
Multi-Agent Interaction Protocol
1. The Contract Net Protocol:
 The most renowned and widely used protocol in multi-agent
systems.
 Describes a protocol involving two kinds of agents:
 (1) a task manager that has to perform a task and does not have the
skills for that, and
 (2) bidders that will perform the task.
 The Contract Net protocol is composed of four phases:
 Task manager Advertises the task to potential bidders (the message also
contains the constraints associated with the task- deadline)
 bidder can bid for the task
 task manager selects one bidder based on its bid and awards it, and
 performing the task.
Multi-Agent Interaction Protocol
 The Contract Net Protocol:
When task execution completes, the awarded bidder
informs the manager that the task is realized (the
inform-done message), or sends the result (the
inform-result message), or sends the failure
message if it fails performing the task.
The first message sent is the cfp (call for proposal)
message, where the task is advertised with its
constraints.
The bidders can answer either by propose
denoting that they agree to realize the task subject
to the constraints associated with it (as expressed
in the message content), or by refuse if the bidder
refuses to execute the task—the reason refusal is
provided by the bidder in the message content
field.
After the deadline, the manager agent selects a bid
based on the constraints.
It then sends an accept proposal to the awarded
bidder and a reject-proposal to all other bidders.
Multi-Agent Interaction Protocol
2. Auction Protocols:
 Many auction protocols are implemented in multi-agent
systems.
 Example: English auction and the Dutch auction
 English auction is an ascending open cry auction where the
price of the item for sale increases as long as at least one
auction participant is willing to pay. It stops when no
participants is willing to raise the price beyond the current
bid.
 Dutch auction protocol is a descending open cry auction.
Unlike the English auction, the price decreases as long as a
participant does not inform it accepts to pay the price.
Multi-Agent Interaction Protocol
The first message inform(start-auction)
informs auction participants that the auction
has begun.
The second message is sent by the auctioneer
(the cfp message).
It informs the auction participants of the starting
bid price.
The participants can respond by a propose
message including a bid equal to or greater
than the current bid.
This can continue iteratively until a deadline is
reached. After a deadline, the auctioneer selects
a bid and informs the winner with the accept-
proposal message.
The other participants who placed bids receive
the reject-proposal message. The auction
continues as long as there are at least two
offers.
The auctioneer may close the auction (the
inform(end of auction)) if there are less than
two offers.
The protocol allows the auctioneer to check whether the hammer price is greater than the
reservation price, and to allocate the sold good to the winner conditioned on this.
Multi-Agent Interaction Protocol
3. Sian’s Learning Protocol:
 A simple protocol where an agent proposes a hypothesis to other
agents. Those agents provide their opinion on whether the
hypothesis is correct or incorrect, and whether they want to modify
it.
 Each agent votes on the correctness of the hypothesis. The agent
that proposed the hypothesis gathers the votes and, using some
heuristics, decides whether the hypothesis is accepted or not. If the
hypothesis is accepted, it informs other agents of the acceptance.
Multi-Agent Interaction Protocol
The first message is a query-if message containing
the hypothesis to validate.
The other agents can answer either by a confirm message
denoting that the hypothesis is correct, by a disagree
message denoting that the hypothesis is wrong, by a
noopinion message denoting that they have no idea about
this hypothesis, or by a modify message requesting to
modify the hypothesis.
Based on heuristics and answers from other agents, the initiator
decides whether the hypothesis is true.
If it decides that the hypothesis is true, it informs the other agents via
an assert message that the hypothesis is considered true.
Dialogue Games
 Conceived with the idea of restoring agents’ autonomy.
 Only pairs of communicative acts are defined, e.g., a request
with an answer.
 Dialogue games are interactions between two or more players,
where each utters according to a defined set of rules.
 Example of dialogues:
 Information-seeking dialogues are those where one
participant seeks the answer to some question(s) from another
participant
 Inquiry dialogues are those where the participants collaborate
to answer some question or questions where answers are not
known to any single participant.
 Persuasion dialogues involve one participant seeking to
persuade another to accept a proposition he or she does not
currently endorse
Dialogue Games
 Example of dialogues:
 Negotiation dialogues are those where the participants
bargain over the division of some scarce resource.
 Deliberation dialogues are those where the participants
collaborate to decide what action or course of action should be
adopted in some situation.
 Eristic dialogues are those where participants quarrel
verbally as a substitute for physical fighting, aiming to vent
perceived grievances.
Dialogue Games: Elements
 Commencement rules: Rules that define the
circumstances under which the dialogue commences.
 Locutions rules: Rules that indicate what utterances
are permitted, and permit those asserting propositions
that are subsequently questioned or contested to justify
their assertions.
 Combination rules: Rules that define the dialogical
contexts under which particular locutions are permitted
or not, or obligatory or not.
 Commitments rules: Rules that define the
circumstances under which participants express
commitment to a proposition.
 Termination rules: Rules that define the circumstances
under which the dialogue ends.
Argumentation Systems
 The previous two Communication assumes what the
agents say to be true.
 Argumentation Systems allow to challenge agents on
their beliefs.
 Agent can use it to, for example, to deal with
inconsistent information.
Multiparty Communication
Proposal
 Communication between agents is private to these agents.
It is then not possible to send the same message to all the
agents and to see the answers without using a multicast or
a broadcast mechanism.
 An agent cannot overhear the communication and intervene
if needed.
 In Internet applications, there are three types of modes
 Public,
 Private and
 Secret
Multiparty Communication Proposal
 Public:
 Resembles the Internet multi-party communication, where potentially all
agents are able to “hear” a message.
 The agent communication not only contains the sender and the addressees
but auditors and over hearers as well.
 Private:
 Private communication is certainly the less frequent mode of communication
in multi-agent systems.
 The aim of this mode is to offer some privacy to agents during the
conversation.
 Agents outside the communication group may be aware of this
communication but cannot perceive the content.
 The aim of this mode of communication is to let agents outside the
communicating group learn that some agents are acquainted.
 Secret:
 Secret communications is similar to current one-to-one communication.
 Communication to be aware of this communication. The approach is the
same as private communications
A purchase negotiation dialogue
game
 The example deals with buying a new car under some
specific constraints such as price, quality, number of
seats, etc.
 The different communicative acts (also called locutions)
in this dialogue game are:
open_dialogue()
enter_dialogue()
seek_info()
willing_to_sell()
desire_to_buy()
prefer()
refuse_to_buy()
refuse_to_sell
agree_to_buy
agree_to_sell
withdraw_dialogue

More Related Content

What's hot

Object Oriented Design in Software Engineering SE12
Object Oriented Design in Software Engineering SE12Object Oriented Design in Software Engineering SE12
Object Oriented Design in Software Engineering SE12
koolkampus
 
Agents in Artificial intelligence
Agents in Artificial intelligence Agents in Artificial intelligence
Agents in Artificial intelligence
Lalit Birla
 
NEGOTIATION AND BARGAINING.pptx
NEGOTIATION AND BARGAINING.pptxNEGOTIATION AND BARGAINING.pptx
NEGOTIATION AND BARGAINING.pptx
THAMARAIKANNANM4
 
Google App Engine
Google App EngineGoogle App Engine
Google App Engine
Software Park Thailand
 
The structure of agents
The structure of agentsThe structure of agents
The structure of agents
Anitha Purushothaman
 
DBMS Assignments Questions
DBMS Assignments QuestionsDBMS Assignments Questions
DBMS Assignments Questions
Sara Sahu
 
T9. Trust and reputation in multi-agent systems
T9. Trust and reputation in multi-agent systemsT9. Trust and reputation in multi-agent systems
T9. Trust and reputation in multi-agent systems
EASSS 2012
 
Multi-agent systems
Multi-agent systemsMulti-agent systems
Multi-agent systems
R A Akerkar
 
LISP: Introduction to lisp
LISP: Introduction to lispLISP: Introduction to lisp
LISP: Introduction to lisp
DataminingTools Inc
 
Distributed Systems Architecture in Software Engineering SE11
Distributed Systems Architecture in Software Engineering SE11Distributed Systems Architecture in Software Engineering SE11
Distributed Systems Architecture in Software Engineering SE11
koolkampus
 
Planning
Planning Planning
Planning
Amar Jukuntla
 
Logics for non monotonic reasoning-ai
Logics for non monotonic reasoning-aiLogics for non monotonic reasoning-ai
Logics for non monotonic reasoning-ai
ShaishavShah8
 
Semantic nets in artificial intelligence
Semantic nets in artificial intelligenceSemantic nets in artificial intelligence
Semantic nets in artificial intelligence
harshita virwani
 
Shared memory
Shared memoryShared memory
Shared memory
Abhishek Khune
 
Uml
UmlUml
Control Strategies in AI
Control Strategies in AI Control Strategies in AI
Control Strategies in AI
Bharat Bhushan
 
AI: Planning and AI
AI: Planning and AIAI: Planning and AI
AI: Planning and AI
DataminingTools Inc
 
Goal based and utility based agents
Goal based and utility based agentsGoal based and utility based agents
Goal based and utility based agents
Megha Sharma
 
Cohesion and coupling
Cohesion and couplingCohesion and coupling
Cohesion and coupling
Aprajita (Abbey) Singh
 
Explain Communication among agents in Artificial Intelligence
Explain Communication among agents in Artificial IntelligenceExplain Communication among agents in Artificial Intelligence
Explain Communication among agents in Artificial Intelligence
Gurpreet singh
 

What's hot (20)

Object Oriented Design in Software Engineering SE12
Object Oriented Design in Software Engineering SE12Object Oriented Design in Software Engineering SE12
Object Oriented Design in Software Engineering SE12
 
Agents in Artificial intelligence
Agents in Artificial intelligence Agents in Artificial intelligence
Agents in Artificial intelligence
 
NEGOTIATION AND BARGAINING.pptx
NEGOTIATION AND BARGAINING.pptxNEGOTIATION AND BARGAINING.pptx
NEGOTIATION AND BARGAINING.pptx
 
Google App Engine
Google App EngineGoogle App Engine
Google App Engine
 
The structure of agents
The structure of agentsThe structure of agents
The structure of agents
 
DBMS Assignments Questions
DBMS Assignments QuestionsDBMS Assignments Questions
DBMS Assignments Questions
 
T9. Trust and reputation in multi-agent systems
T9. Trust and reputation in multi-agent systemsT9. Trust and reputation in multi-agent systems
T9. Trust and reputation in multi-agent systems
 
Multi-agent systems
Multi-agent systemsMulti-agent systems
Multi-agent systems
 
LISP: Introduction to lisp
LISP: Introduction to lispLISP: Introduction to lisp
LISP: Introduction to lisp
 
Distributed Systems Architecture in Software Engineering SE11
Distributed Systems Architecture in Software Engineering SE11Distributed Systems Architecture in Software Engineering SE11
Distributed Systems Architecture in Software Engineering SE11
 
Planning
Planning Planning
Planning
 
Logics for non monotonic reasoning-ai
Logics for non monotonic reasoning-aiLogics for non monotonic reasoning-ai
Logics for non monotonic reasoning-ai
 
Semantic nets in artificial intelligence
Semantic nets in artificial intelligenceSemantic nets in artificial intelligence
Semantic nets in artificial intelligence
 
Shared memory
Shared memoryShared memory
Shared memory
 
Uml
UmlUml
Uml
 
Control Strategies in AI
Control Strategies in AI Control Strategies in AI
Control Strategies in AI
 
AI: Planning and AI
AI: Planning and AIAI: Planning and AI
AI: Planning and AI
 
Goal based and utility based agents
Goal based and utility based agentsGoal based and utility based agents
Goal based and utility based agents
 
Cohesion and coupling
Cohesion and couplingCohesion and coupling
Cohesion and coupling
 
Explain Communication among agents in Artificial Intelligence
Explain Communication among agents in Artificial IntelligenceExplain Communication among agents in Artificial Intelligence
Explain Communication among agents in Artificial Intelligence
 

Viewers also liked

Multiagent System Communication
Multiagent System Communication Multiagent System Communication
Multiagent System Communication
Ahsan Rahim
 
|.doc|
|.doc||.doc|
|.doc|
butest
 
Topic 1 lecture 2
Topic 1 lecture 2Topic 1 lecture 2
Topic 1 lecture 2
farshad33
 
Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...
Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...
Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...
farshad33
 
Topic 1 lecture 3-application imapct of mas&t
Topic 1 lecture 3-application imapct of mas&tTopic 1 lecture 3-application imapct of mas&t
Topic 1 lecture 3-application imapct of mas&t
farshad33
 
Multiagent systems (and their use in industry)
Multiagent systems (and their use in industry)Multiagent systems (and their use in industry)
Multiagent systems (and their use in industry)
Marc-Philippe Huget
 
Chapter 5 design patterns for mas
Chapter 5 design patterns for masChapter 5 design patterns for mas
Chapter 5 design patterns for mas
farshad33
 
Auctions
AuctionsAuctions
Auctions
Mukesh Kumar
 
Topic 1 lecture 1
Topic 1 lecture 1Topic 1 lecture 1
Topic 1 lecture 1
farshad33
 
Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...
Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...
Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...
farshad33
 
Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...
Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...
Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...
farshad33
 
Introduction to Agents and Multi-agent Systems (lecture slides)
Introduction to Agents and Multi-agent Systems (lecture slides)Introduction to Agents and Multi-agent Systems (lecture slides)
Introduction to Agents and Multi-agent Systems (lecture slides)
Dagmar Monett
 
Introduction to agents and multi-agent systems
Introduction to agents and multi-agent systemsIntroduction to agents and multi-agent systems
Introduction to agents and multi-agent systems
Antonio Moreno
 

Viewers also liked (13)

Multiagent System Communication
Multiagent System Communication Multiagent System Communication
Multiagent System Communication
 
|.doc|
|.doc||.doc|
|.doc|
 
Topic 1 lecture 2
Topic 1 lecture 2Topic 1 lecture 2
Topic 1 lecture 2
 
Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...
Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...
Chapter 8 agent-oriented software engineering ch8-prometheus research methodo...
 
Topic 1 lecture 3-application imapct of mas&t
Topic 1 lecture 3-application imapct of mas&tTopic 1 lecture 3-application imapct of mas&t
Topic 1 lecture 3-application imapct of mas&t
 
Multiagent systems (and their use in industry)
Multiagent systems (and their use in industry)Multiagent systems (and their use in industry)
Multiagent systems (and their use in industry)
 
Chapter 5 design patterns for mas
Chapter 5 design patterns for masChapter 5 design patterns for mas
Chapter 5 design patterns for mas
 
Auctions
AuctionsAuctions
Auctions
 
Topic 1 lecture 1
Topic 1 lecture 1Topic 1 lecture 1
Topic 1 lecture 1
 
Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...
Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...
Chapter 7 agent-oriented software engineering ch7-agent methodology-agent met...
 
Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...
Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...
Topic 4 -software architecture viewpoint-multi-agent systems-a software archi...
 
Introduction to Agents and Multi-agent Systems (lecture slides)
Introduction to Agents and Multi-agent Systems (lecture slides)Introduction to Agents and Multi-agent Systems (lecture slides)
Introduction to Agents and Multi-agent Systems (lecture slides)
 
Introduction to agents and multi-agent systems
Introduction to agents and multi-agent systemsIntroduction to agents and multi-agent systems
Introduction to agents and multi-agent systems
 

Similar to Chapter 6 agent communications--agent communications

What kinds of languages can agents use to communicate?
What kinds of languages can agents use to communicate?What kinds of languages can agents use to communicate?
What kinds of languages can agents use to communicate?
Ahsan Rahim
 
All about agents jade
All about agents jadeAll about agents jade
All about agents jade
Aryan Rathore
 
Assessment of KQML Improved
Assessment of KQML ImprovedAssessment of KQML Improved
Assessment of KQML Improved
Eswar Publications
 
@lis agent communication, ontologies, protocols, semantic web 2003
@lis   agent communication, ontologies, protocols, semantic web 2003@lis   agent communication, ontologies, protocols, semantic web 2003
@lis agent communication, ontologies, protocols, semantic web 2003
Luigi Ceccaroni
 
agent uml الوكيل باستخدام لغة النمذجة الموحدة
  agent uml الوكيل باستخدام لغة النمذجة الموحدة  agent uml الوكيل باستخدام لغة النمذجة الموحدة
agent uml الوكيل باستخدام لغة النمذجة الموحدة
Mohamed Elagnaf
 
CSCorganization of programming languages
CSCorganization of programming languagesCSCorganization of programming languages
CSCorganization of programming languages
OluwafolakeOjo
 
AN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGE
AN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGEAN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGE
AN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGE
IJwest
 
An Intersemiotic Translation of Normative Utterances to Machine Language
An Intersemiotic Translation of Normative Utterances to Machine LanguageAn Intersemiotic Translation of Normative Utterances to Machine Language
An Intersemiotic Translation of Normative Utterances to Machine Language
dannyijwest
 
Cooperation and Collaboration in Natural Command Language Dialogue.docx
Cooperation and Collaboration in Natural Command Language Dialogue.docxCooperation and Collaboration in Natural Command Language Dialogue.docx
Cooperation and Collaboration in Natural Command Language Dialogue.docx
donnajames55
 
Core idea driving and defining OO: using dynamic polymorphism to invert key a...
Core idea driving and defining OO: using dynamic polymorphism to invert key a...Core idea driving and defining OO: using dynamic polymorphism to invert key a...
Core idea driving and defining OO: using dynamic polymorphism to invert key a...
Philip Schwarz
 
Combining the functional and oject oriented paradigms in the fobs-x scripting...
Combining the functional and oject oriented paradigms in the fobs-x scripting...Combining the functional and oject oriented paradigms in the fobs-x scripting...
Combining the functional and oject oriented paradigms in the fobs-x scripting...
ijpla
 
STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...
STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...
STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...
ijnlc
 
54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx
54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx
54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx
blondellchancy
 
**JUNK** (no subject)
**JUNK** (no subject)**JUNK** (no subject)
**JUNK** (no subject)
muthukumaran.tdr95
 
Principles of Health Informatics: Communication systems in healthcare
Principles of Health Informatics: Communication systems in healthcarePrinciples of Health Informatics: Communication systems in healthcare
Principles of Health Informatics: Communication systems in healthcare
Martin Chapman
 
Html
HtmlHtml
Html
lidanx
 
Suitability of UML State Machine for Modeling Choreography of Services
Suitability of UML State Machine for Modeling Choreography of Services  Suitability of UML State Machine for Modeling Choreography of Services
Suitability of UML State Machine for Modeling Choreography of Services
dannyijwest
 
Machine translation with statistical approach
Machine translation with statistical approachMachine translation with statistical approach
Machine translation with statistical approach
vini89
 
NLP_A Chat-Bot_answering_queries_of_UT-Dallas_Students
NLP_A Chat-Bot_answering_queries_of_UT-Dallas_StudentsNLP_A Chat-Bot_answering_queries_of_UT-Dallas_Students
NLP_A Chat-Bot_answering_queries_of_UT-Dallas_Students
Himanshu kandwal
 
Cyber Language Analysis: Emerging Linguist Trends in Digital Environments
Cyber Language Analysis: Emerging Linguist Trends in Digital EnvironmentsCyber Language Analysis: Emerging Linguist Trends in Digital Environments
Cyber Language Analysis: Emerging Linguist Trends in Digital Environments
Lingua Brava
 

Similar to Chapter 6 agent communications--agent communications (20)

What kinds of languages can agents use to communicate?
What kinds of languages can agents use to communicate?What kinds of languages can agents use to communicate?
What kinds of languages can agents use to communicate?
 
All about agents jade
All about agents jadeAll about agents jade
All about agents jade
 
Assessment of KQML Improved
Assessment of KQML ImprovedAssessment of KQML Improved
Assessment of KQML Improved
 
@lis agent communication, ontologies, protocols, semantic web 2003
@lis   agent communication, ontologies, protocols, semantic web 2003@lis   agent communication, ontologies, protocols, semantic web 2003
@lis agent communication, ontologies, protocols, semantic web 2003
 
agent uml الوكيل باستخدام لغة النمذجة الموحدة
  agent uml الوكيل باستخدام لغة النمذجة الموحدة  agent uml الوكيل باستخدام لغة النمذجة الموحدة
agent uml الوكيل باستخدام لغة النمذجة الموحدة
 
CSCorganization of programming languages
CSCorganization of programming languagesCSCorganization of programming languages
CSCorganization of programming languages
 
AN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGE
AN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGEAN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGE
AN INTERSEMIOTIC TRANSLATION OF NORMATIVE UTTERANCES TO MACHINE LANGUAGE
 
An Intersemiotic Translation of Normative Utterances to Machine Language
An Intersemiotic Translation of Normative Utterances to Machine LanguageAn Intersemiotic Translation of Normative Utterances to Machine Language
An Intersemiotic Translation of Normative Utterances to Machine Language
 
Cooperation and Collaboration in Natural Command Language Dialogue.docx
Cooperation and Collaboration in Natural Command Language Dialogue.docxCooperation and Collaboration in Natural Command Language Dialogue.docx
Cooperation and Collaboration in Natural Command Language Dialogue.docx
 
Core idea driving and defining OO: using dynamic polymorphism to invert key a...
Core idea driving and defining OO: using dynamic polymorphism to invert key a...Core idea driving and defining OO: using dynamic polymorphism to invert key a...
Core idea driving and defining OO: using dynamic polymorphism to invert key a...
 
Combining the functional and oject oriented paradigms in the fobs-x scripting...
Combining the functional and oject oriented paradigms in the fobs-x scripting...Combining the functional and oject oriented paradigms in the fobs-x scripting...
Combining the functional and oject oriented paradigms in the fobs-x scripting...
 
STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...
STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...
STRUCTURED AND QUANTITATIVE PROPERTIES OF ARABIC SMS-BASED CLASSIFIED ADS SUB...
 
54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx
54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx
54-6 September 2002, Edinburgh, UK, Pages 5-11.Bos, Fost.docx
 
**JUNK** (no subject)
**JUNK** (no subject)**JUNK** (no subject)
**JUNK** (no subject)
 
Principles of Health Informatics: Communication systems in healthcare
Principles of Health Informatics: Communication systems in healthcarePrinciples of Health Informatics: Communication systems in healthcare
Principles of Health Informatics: Communication systems in healthcare
 
Html
HtmlHtml
Html
 
Suitability of UML State Machine for Modeling Choreography of Services
Suitability of UML State Machine for Modeling Choreography of Services  Suitability of UML State Machine for Modeling Choreography of Services
Suitability of UML State Machine for Modeling Choreography of Services
 
Machine translation with statistical approach
Machine translation with statistical approachMachine translation with statistical approach
Machine translation with statistical approach
 
NLP_A Chat-Bot_answering_queries_of_UT-Dallas_Students
NLP_A Chat-Bot_answering_queries_of_UT-Dallas_StudentsNLP_A Chat-Bot_answering_queries_of_UT-Dallas_Students
NLP_A Chat-Bot_answering_queries_of_UT-Dallas_Students
 
Cyber Language Analysis: Emerging Linguist Trends in Digital Environments
Cyber Language Analysis: Emerging Linguist Trends in Digital EnvironmentsCyber Language Analysis: Emerging Linguist Trends in Digital Environments
Cyber Language Analysis: Emerging Linguist Trends in Digital Environments
 

Recently uploaded

How to Add a Filter in the Odoo 17 - Odoo 17 Slides
How to Add a Filter in the Odoo 17 - Odoo 17 SlidesHow to Add a Filter in the Odoo 17 - Odoo 17 Slides
How to Add a Filter in the Odoo 17 - Odoo 17 Slides
Celine George
 
Imagination in Computer Science Research
Imagination in Computer Science ResearchImagination in Computer Science Research
Imagination in Computer Science Research
Abhik Roychoudhury
 
Mail Server Configuration Using App passwords in Odoo 17
Mail Server Configuration Using App passwords in Odoo 17Mail Server Configuration Using App passwords in Odoo 17
Mail Server Configuration Using App passwords in Odoo 17
Celine George
 
Open Source and AI - ByWater Closing Keynote Presentation.pdf
Open Source and AI - ByWater Closing Keynote Presentation.pdfOpen Source and AI - ByWater Closing Keynote Presentation.pdf
Open Source and AI - ByWater Closing Keynote Presentation.pdf
Jessica Zairo
 
SD_Integrating 21st Century Skills in Classroom-based Assessment.pptx
SD_Integrating 21st Century Skills in Classroom-based Assessment.pptxSD_Integrating 21st Century Skills in Classroom-based Assessment.pptx
SD_Integrating 21st Century Skills in Classroom-based Assessment.pptx
elwoodprias1
 
E-learning Odoo 17 New features - Odoo 17 Slides
E-learning Odoo 17  New features - Odoo 17 SlidesE-learning Odoo 17  New features - Odoo 17 Slides
E-learning Odoo 17 New features - Odoo 17 Slides
Celine George
 
Demonstration module in Odoo 17 - Odoo 17 Slides
Demonstration module in Odoo 17 - Odoo 17 SlidesDemonstration module in Odoo 17 - Odoo 17 Slides
Demonstration module in Odoo 17 - Odoo 17 Slides
Celine George
 
New Features in Odoo 17 Sign - Odoo 17 Slides
New Features in Odoo 17 Sign - Odoo 17 SlidesNew Features in Odoo 17 Sign - Odoo 17 Slides
New Features in Odoo 17 Sign - Odoo 17 Slides
Celine George
 
How to Manage Early Receipt Printing in Odoo 17 POS
How to Manage Early Receipt Printing in Odoo 17 POSHow to Manage Early Receipt Printing in Odoo 17 POS
How to Manage Early Receipt Printing in Odoo 17 POS
Celine George
 
A beginner’s guide to project reviews - everything you wanted to know but wer...
A beginner’s guide to project reviews - everything you wanted to know but wer...A beginner’s guide to project reviews - everything you wanted to know but wer...
A beginner’s guide to project reviews - everything you wanted to know but wer...
Association for Project Management
 
QCE – Unpacking the syllabus Implications for Senior School practices and ass...
QCE – Unpacking the syllabus Implications for Senior School practices and ass...QCE – Unpacking the syllabus Implications for Senior School practices and ass...
QCE – Unpacking the syllabus Implications for Senior School practices and ass...
mansk2
 
Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)
Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)
Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)
Cátedra Banco Santander
 
Node JS Interview Question PDF By ScholarHat
Node JS Interview Question PDF By ScholarHatNode JS Interview Question PDF By ScholarHat
Node JS Interview Question PDF By ScholarHat
Scholarhat
 
RDBMS Lecture Notes Unit4 chapter12 VIEW
RDBMS Lecture Notes Unit4 chapter12 VIEWRDBMS Lecture Notes Unit4 chapter12 VIEW
RDBMS Lecture Notes Unit4 chapter12 VIEW
Murugan Solaiyappan
 
C Interview Questions PDF By Scholarhat.pdf
C Interview Questions PDF By Scholarhat.pdfC Interview Questions PDF By Scholarhat.pdf
C Interview Questions PDF By Scholarhat.pdf
Scholarhat
 
2 Post harvest Physiology of Horticulture produce.pptx
2 Post harvest Physiology of Horticulture  produce.pptx2 Post harvest Physiology of Horticulture  produce.pptx
2 Post harvest Physiology of Horticulture produce.pptx
UmeshTimilsina1
 
ASP.NET Core Interview Questions PDF By ScholarHat.pdf
ASP.NET Core Interview Questions PDF By ScholarHat.pdfASP.NET Core Interview Questions PDF By ScholarHat.pdf
ASP.NET Core Interview Questions PDF By ScholarHat.pdf
Scholarhat
 
Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...
Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...
Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...
Alvaro Barbosa
 
PRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdf
PRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdfPRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdf
PRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdf
nservice241
 
Parent PD Design for Professional Development .docx
Parent PD Design for Professional Development .docxParent PD Design for Professional Development .docx
Parent PD Design for Professional Development .docx
AntonioJarligoCompra
 

Recently uploaded (20)

How to Add a Filter in the Odoo 17 - Odoo 17 Slides
How to Add a Filter in the Odoo 17 - Odoo 17 SlidesHow to Add a Filter in the Odoo 17 - Odoo 17 Slides
How to Add a Filter in the Odoo 17 - Odoo 17 Slides
 
Imagination in Computer Science Research
Imagination in Computer Science ResearchImagination in Computer Science Research
Imagination in Computer Science Research
 
Mail Server Configuration Using App passwords in Odoo 17
Mail Server Configuration Using App passwords in Odoo 17Mail Server Configuration Using App passwords in Odoo 17
Mail Server Configuration Using App passwords in Odoo 17
 
Open Source and AI - ByWater Closing Keynote Presentation.pdf
Open Source and AI - ByWater Closing Keynote Presentation.pdfOpen Source and AI - ByWater Closing Keynote Presentation.pdf
Open Source and AI - ByWater Closing Keynote Presentation.pdf
 
SD_Integrating 21st Century Skills in Classroom-based Assessment.pptx
SD_Integrating 21st Century Skills in Classroom-based Assessment.pptxSD_Integrating 21st Century Skills in Classroom-based Assessment.pptx
SD_Integrating 21st Century Skills in Classroom-based Assessment.pptx
 
E-learning Odoo 17 New features - Odoo 17 Slides
E-learning Odoo 17  New features - Odoo 17 SlidesE-learning Odoo 17  New features - Odoo 17 Slides
E-learning Odoo 17 New features - Odoo 17 Slides
 
Demonstration module in Odoo 17 - Odoo 17 Slides
Demonstration module in Odoo 17 - Odoo 17 SlidesDemonstration module in Odoo 17 - Odoo 17 Slides
Demonstration module in Odoo 17 - Odoo 17 Slides
 
New Features in Odoo 17 Sign - Odoo 17 Slides
New Features in Odoo 17 Sign - Odoo 17 SlidesNew Features in Odoo 17 Sign - Odoo 17 Slides
New Features in Odoo 17 Sign - Odoo 17 Slides
 
How to Manage Early Receipt Printing in Odoo 17 POS
How to Manage Early Receipt Printing in Odoo 17 POSHow to Manage Early Receipt Printing in Odoo 17 POS
How to Manage Early Receipt Printing in Odoo 17 POS
 
A beginner’s guide to project reviews - everything you wanted to know but wer...
A beginner’s guide to project reviews - everything you wanted to know but wer...A beginner’s guide to project reviews - everything you wanted to know but wer...
A beginner’s guide to project reviews - everything you wanted to know but wer...
 
QCE – Unpacking the syllabus Implications for Senior School practices and ass...
QCE – Unpacking the syllabus Implications for Senior School practices and ass...QCE – Unpacking the syllabus Implications for Senior School practices and ass...
QCE – Unpacking the syllabus Implications for Senior School practices and ass...
 
Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)
Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)
Cómo crear video-tutoriales con ScreenPal (2 de julio de 2024)
 
Node JS Interview Question PDF By ScholarHat
Node JS Interview Question PDF By ScholarHatNode JS Interview Question PDF By ScholarHat
Node JS Interview Question PDF By ScholarHat
 
RDBMS Lecture Notes Unit4 chapter12 VIEW
RDBMS Lecture Notes Unit4 chapter12 VIEWRDBMS Lecture Notes Unit4 chapter12 VIEW
RDBMS Lecture Notes Unit4 chapter12 VIEW
 
C Interview Questions PDF By Scholarhat.pdf
C Interview Questions PDF By Scholarhat.pdfC Interview Questions PDF By Scholarhat.pdf
C Interview Questions PDF By Scholarhat.pdf
 
2 Post harvest Physiology of Horticulture produce.pptx
2 Post harvest Physiology of Horticulture  produce.pptx2 Post harvest Physiology of Horticulture  produce.pptx
2 Post harvest Physiology of Horticulture produce.pptx
 
ASP.NET Core Interview Questions PDF By ScholarHat.pdf
ASP.NET Core Interview Questions PDF By ScholarHat.pdfASP.NET Core Interview Questions PDF By ScholarHat.pdf
ASP.NET Core Interview Questions PDF By ScholarHat.pdf
 
Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...
Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...
Benchmarking Sustainability: Neurosciences and AI Tech Research in Macau - Ke...
 
PRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdf
PRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdfPRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdf
PRESS RELEASE - UNIVERSITY OF GHANA, JULY 16, 2024.pdf
 
Parent PD Design for Professional Development .docx
Parent PD Design for Professional Development .docxParent PD Design for Professional Development .docx
Parent PD Design for Professional Development .docx
 

Chapter 6 agent communications--agent communications

  • 2. Introduction  Agent communication is essential for facilitating cooperation, coordination, and other aspects of agent interaction  Two major approaches to agent communication  Direct communication, and  Indirect communication  Early work on agent communication focused on direct communication with the use of different protocols  Protocols are efficient means for information exchange and coordination; however, they restrict agents’ autonomy; agents have to strictly follow the admissible sequence of messages imposed by the protocol.
  • 3. Introduction  Agent communication is based on human conversations.  Message contains a verb and a content.  For instance: inform (it rains) inform is the verb and it rains is the content. F(P): F- force; P- message content  Several types of speech acts exist in human conversations:  Representatives, which give information  Directives, which require something from the recipient  Commissives, which commit the sender to do something in the future  Expressives, which describe the mental states of the sender  Declarations, which process an act just by saying it  Agents do not use all the types of verbs present in a natural language but a smaller subset.
  • 4. Direct Communication  Communication between cognitive agents (agents that have a representation of themselves, of other agents, and of the world, and reason about their courses of action) where the recipient is formally defined  Several elements require attention when considering direct communication:  an agent communication language  ontologies  communication support
  • 5. Indirect Communication  Message is released in the environment. Any agent sufficiently close to it can perceive it.  Indirect communication is used by reactive agents (agents that do not have a representation of themselves, of other agents, and of the world.  Can only be used for limited coordination, since it is not possible to represent a wide range of messages.  Examples of indirect communication include pheromones (example the MANTA project ) and forces (AI, robotics, and multi-agent systems).
  • 6. Agent Communication Versus Multi-agent Communication  Agent communication (as opposed to multi-agent communication) is a one-to-one communication in which a speaker utters speech acts to a single addressee.  Several features characterize one-to-one communication:  Communication is between two agents: a speaker and an addressee.  Communication is privileged between two agents, that is, it is not possible for other agents to hear the communication.  Entering and leaving communications are subject to agreement from other  Terms of termination of communication are necessarily known before beginning the communication.
  • 7. Agent Communication Versus Multi-agent Communication  Multi-agent communication allow for indirect communication. That is, it should allow agents to listen to messages transmitted in their vicinity (either physical or logical) but not too far-away communication.  It is then important to add the notion of earshot. In human communication, earshot can range from whispering (private communications) to shouting where even further agents can hear.  Multi-agent systems are open systems; as a consequence, agents can enter and leave at any moment during the conversation.  Termination of communications is not necessarily decided in advance. Termination can happen if no more communication is needed.
  • 8. Direct Communication: Agent Communication Languages  Communicating requires the use of common language. For example, we communicate using English, to understand each other.  Agent communication languages do not offer a rich set of communicative acts as natural languages.  Currently, there exist two main agent communication languages: KQML (Knowledge Query and Manipulation Language) and FIPA ACL.  KQML allow data exchange between distributed knowledge bases. KQML is composed of  7 performatives for conversation management (sorry, error, etc.),  17 performatives for the conversation (ask-if, tell, etc.), and  11 performatives for conversation sending (forward, broker-all, etc.).
  • 10. Direct Communication: Agent Communication Languages  KQML is the older communication languages and is gradually replaced by FIPA ACL for several reasons:  There was a need for clear semantics for the language, (make a clear different between KQML and FIPA ACL)  It was recommended to have a smaller set of communicative acts, while keeping the autonomy of agents,  FIPA ACL has implementation of various protocols, and  FIPA ACL supports communication at various levels.
  • 11. The KQML Language  Knowledge Query and Manipulation Language (KQML) is a language that is designed to support interaction among intelligent software agents.  It was developed by the ARPA-supported Knowledge Sharing Effort and independently implemented by several research groups.  It has been successfully used to implement a variety of information systems using different software architectures.  The KQML language simplifies its implementation by allowing KQML messages to carry arbitrary useful information, such as the names and addresses of the sending and receiving agents, a unique message identifier, and notations by any intervening agents.
  • 12. KQML: the 3 layers The KQML language can be viewed as consisting of three layers: the content, message and communication layers, as shown in Figure. • The content layer specifies the actual content of the message, and the KQML standard itself has nothing to say about this. • The set of performatives provided by the language, and shown in Table, constitute the message layer, which in turn forms the core of the language. • This layer of abstraction provides the performative and specifies the protocol for delivering the message that subsumes the content. • KQML specifies several protocols, including synchronous (where a blocking query waits for an expected reply) and asynchronous (which involves non- blocking messages). • The communication layer encodes low level communication parameters, such as the identities of the sender and the recipient, and unique identifiers for the particular speech act.
  • 13. Example of an instantiated performative A KQML message from agent joe representing a query about the price of a share of IBM stock might be encoded as: In this message, the KQML performative is ask-one, the content is (price ibm ?price), the ontology assumed by the query is identified by the token nyse-ticks, the receiver of the message is to be a server identified as stock-server and the query is written in a language called LPROLOG. The value of the :content keyword is the content level, the values of the :reply-with, :sender, :receiver keywords form the communication layer and the performative name, with the :language and :ontology form the message layer.
  • 14. Applications of KQML Language  KQML has been used as the communication language in several technology integration experiments in the ARPA Rome Lab Planning Initiative.  One of these experiments supported an integrated planning and scheduling system for military transportation logistics linking a planning agent, with a scheduler (in Common Lisp), a knowledge base, and a case based reasoning tool (in Common Lisp).  All of the components integrated were pre-existing systems which were not designed to work in a cooperative distributed environment.
  • 15. FIPA ACL  The Foundation for Intelligent Physical Agents (FIPA) was formed in 1996 to produce software standards for heterogeneous and interacting agents and agent-based systems.  FIPA’s agent communication language, like KQML, is based on speech act theory: messages are actions or communicative acts, as they are intended to perform some action by virtue of being sent.  FIPA ACL is superficially similar to KQML. Its syntax is identical to KQML’s except for different names for some reserved primitives.  Thus, it maintains the KQML approach of separating the outer language from the inner language. The outer language defines the intended meaning of the message; the inner, or content, language denotes the variety of interesting agent architectures.  Thus, KQML introduces a small number of KQML performatives that agents use to describe the metadata specifying the information requirements and capabilities; it also introduces a special class of agents called communication facilitators.  A facilitator is an agent that performs various useful communication services, such as maintaining a registry of service names, forwarding messages to named services, routing messages based on content, matchmaking between information providers and clients, and providing mediation and translation services.
  • 16. Agent Communication  This chapter will describe FIPA ACL.  F(P); where F is force P is message content  For example: inform (it rains); Force (F) is the performative verb  Performatives in FIPA ACL are arranged into five categories:  Passing information,  Requesting information,  Negotiation,  Performing actions, and  Error handling.  However, we note that agent-based applications can be (and have been) developed using traditional third generation languages like Lisp, C, or Prolog, and object-oriented (OO) languages such as C++ or Smalltalk.
  • 17. Direct Communication: Agent Communication Languages Primitive performatives
  • 18. FIPA ACL  Inform: an agent informs recipients about something it believes and it believes recipients do not yet believe it.  Request: request that recipient perform actions.  KQML is lack of semantics associated with performative. FIPA ACL addresses this issue via two parts:  Feasible pre condition: pre condition that have to be evaluated to true to execute the performative.  Rational effect: expresses what the sender hopes to bring about.  Semantics are expresses via FIPA SL (refer table 6.2)
  • 19. Direct Communication: Ontologies  It is possible that two agents speak about the same concept but not with the same term: using the terms “painter” and “artist,” for instance.  In order to avoid these differences, it is important to use a common vocabulary also called an ontology  An ontology defines the basic terms and relations comprising the vocabulary of a topic area as well as the rules for combining terms and relations to define extensions to the vocabulary  Example of ontology for computer science department is shown next:
  • 21. Direct Communication: Communication Support  Communicating requires a communication support, which is a mechanism that stores, retrieves, and directs messages to the agents.  Communication support is present in agent platforms such as JADE and MadKit , where it is possible to send messages to  Agents,  The roles they play in a group, or  A group.
  • 22. Direct Communication: Structuring Communication Three main different approaches exist to structure direct communication: 1. Protocols: Main approach to represent communication in MAS systems. The way to structure communication in a rigid manner. A finite set of performatives is allowed for a particular state of the interaction. 2. Dialogue games: offer more flexibility in communication in comparison with protocols. The communicative acts are paired, for instance, a question is paired with answer. It is possible to embed dialogue games into other dialogue games. 3. Argumentation systems: agents challenge the content of each message and as a consequence, agents have to justify their message content via arguments. Argumentation is frequently used in negotiation-based systems.
  • 23. Protocol  Research on protocols in multi-agent systems share work with communication protocol engineering and interaction protocol engineering.  Interaction protocol engineering composed of several phases from analysis to execution
  • 24. Interaction Protocol Engineering 1. Analysis: This phase describes informally the protocol. The exact sequence of messages, the content, and the different requirements the protocol should follow. 2. Formal Description : To avoid misunderstandings and to ease validation. Several formal description techniques are available for interaction protocol designers:  Finite State Machines : Based on graph theory, where each state corresponds to an interaction state and a transition sends or receives a message. Example include AgenTalk , and COOL  Petri nets: Allow multiple flows and synchronization on the same net.  UML-based notations: Example FIPA UAML and Agent UML  Others: temporal logic and Z language
  • 25. Interaction Protocol Engineering 3. Validation: The validation can be structural (checking for deadlock, termination, etc.) or behavioral (checking whether the protocol achieves its goals). Two techniques exist for the validation of interaction protocols:  Reachability analysis : A graph of all the reachable states of the protocol is built-validation is based on graph properties.  Model checking: A graph of reachable states is built, but this time the property is expressed as a temporal logic formula.  Another kind of validation is the agent communication language compliance verification, which checks whether an agent correctly implements the protocol and agent communication language.
  • 26. Interaction Protocol Engineering 4. Protocol Synthesis: The actual implementation. First the skeleton of the protocol is generated. Then the semantics associated with each transition is added. The first phase is automatic and the second one is manual. System unable to generate the semantics. The advantage is that it reduces the effort required for generating a protocol since it s not necessary to rewrite each time the semantics of the protocol modified. 5. Conformance Testing: Checks whether the implemented protocol contains the requirements defined in the analysis phase.
  • 27. Multi-Agent Interaction Protocol 1. The Contract Net Protocol:  The most renowned and widely used protocol in multi-agent systems.  Describes a protocol involving two kinds of agents:  (1) a task manager that has to perform a task and does not have the skills for that, and  (2) bidders that will perform the task.  The Contract Net protocol is composed of four phases:  Task manager Advertises the task to potential bidders (the message also contains the constraints associated with the task- deadline)  bidder can bid for the task  task manager selects one bidder based on its bid and awards it, and  performing the task.
  • 28. Multi-Agent Interaction Protocol  The Contract Net Protocol: When task execution completes, the awarded bidder informs the manager that the task is realized (the inform-done message), or sends the result (the inform-result message), or sends the failure message if it fails performing the task. The first message sent is the cfp (call for proposal) message, where the task is advertised with its constraints. The bidders can answer either by propose denoting that they agree to realize the task subject to the constraints associated with it (as expressed in the message content), or by refuse if the bidder refuses to execute the task—the reason refusal is provided by the bidder in the message content field. After the deadline, the manager agent selects a bid based on the constraints. It then sends an accept proposal to the awarded bidder and a reject-proposal to all other bidders.
  • 29. Multi-Agent Interaction Protocol 2. Auction Protocols:  Many auction protocols are implemented in multi-agent systems.  Example: English auction and the Dutch auction  English auction is an ascending open cry auction where the price of the item for sale increases as long as at least one auction participant is willing to pay. It stops when no participants is willing to raise the price beyond the current bid.  Dutch auction protocol is a descending open cry auction. Unlike the English auction, the price decreases as long as a participant does not inform it accepts to pay the price.
  • 30. Multi-Agent Interaction Protocol The first message inform(start-auction) informs auction participants that the auction has begun. The second message is sent by the auctioneer (the cfp message). It informs the auction participants of the starting bid price. The participants can respond by a propose message including a bid equal to or greater than the current bid. This can continue iteratively until a deadline is reached. After a deadline, the auctioneer selects a bid and informs the winner with the accept- proposal message. The other participants who placed bids receive the reject-proposal message. The auction continues as long as there are at least two offers. The auctioneer may close the auction (the inform(end of auction)) if there are less than two offers. The protocol allows the auctioneer to check whether the hammer price is greater than the reservation price, and to allocate the sold good to the winner conditioned on this.
  • 31. Multi-Agent Interaction Protocol 3. Sian’s Learning Protocol:  A simple protocol where an agent proposes a hypothesis to other agents. Those agents provide their opinion on whether the hypothesis is correct or incorrect, and whether they want to modify it.  Each agent votes on the correctness of the hypothesis. The agent that proposed the hypothesis gathers the votes and, using some heuristics, decides whether the hypothesis is accepted or not. If the hypothesis is accepted, it informs other agents of the acceptance.
  • 32. Multi-Agent Interaction Protocol The first message is a query-if message containing the hypothesis to validate. The other agents can answer either by a confirm message denoting that the hypothesis is correct, by a disagree message denoting that the hypothesis is wrong, by a noopinion message denoting that they have no idea about this hypothesis, or by a modify message requesting to modify the hypothesis. Based on heuristics and answers from other agents, the initiator decides whether the hypothesis is true. If it decides that the hypothesis is true, it informs the other agents via an assert message that the hypothesis is considered true.
  • 33. Dialogue Games  Conceived with the idea of restoring agents’ autonomy.  Only pairs of communicative acts are defined, e.g., a request with an answer.  Dialogue games are interactions between two or more players, where each utters according to a defined set of rules.  Example of dialogues:  Information-seeking dialogues are those where one participant seeks the answer to some question(s) from another participant  Inquiry dialogues are those where the participants collaborate to answer some question or questions where answers are not known to any single participant.  Persuasion dialogues involve one participant seeking to persuade another to accept a proposition he or she does not currently endorse
  • 34. Dialogue Games  Example of dialogues:  Negotiation dialogues are those where the participants bargain over the division of some scarce resource.  Deliberation dialogues are those where the participants collaborate to decide what action or course of action should be adopted in some situation.  Eristic dialogues are those where participants quarrel verbally as a substitute for physical fighting, aiming to vent perceived grievances.
  • 35. Dialogue Games: Elements  Commencement rules: Rules that define the circumstances under which the dialogue commences.  Locutions rules: Rules that indicate what utterances are permitted, and permit those asserting propositions that are subsequently questioned or contested to justify their assertions.  Combination rules: Rules that define the dialogical contexts under which particular locutions are permitted or not, or obligatory or not.  Commitments rules: Rules that define the circumstances under which participants express commitment to a proposition.  Termination rules: Rules that define the circumstances under which the dialogue ends.
  • 36. Argumentation Systems  The previous two Communication assumes what the agents say to be true.  Argumentation Systems allow to challenge agents on their beliefs.  Agent can use it to, for example, to deal with inconsistent information.
  • 37. Multiparty Communication Proposal  Communication between agents is private to these agents. It is then not possible to send the same message to all the agents and to see the answers without using a multicast or a broadcast mechanism.  An agent cannot overhear the communication and intervene if needed.  In Internet applications, there are three types of modes  Public,  Private and  Secret
  • 38. Multiparty Communication Proposal  Public:  Resembles the Internet multi-party communication, where potentially all agents are able to “hear” a message.  The agent communication not only contains the sender and the addressees but auditors and over hearers as well.  Private:  Private communication is certainly the less frequent mode of communication in multi-agent systems.  The aim of this mode is to offer some privacy to agents during the conversation.  Agents outside the communication group may be aware of this communication but cannot perceive the content.  The aim of this mode of communication is to let agents outside the communicating group learn that some agents are acquainted.  Secret:  Secret communications is similar to current one-to-one communication.  Communication to be aware of this communication. The approach is the same as private communications
  • 39. A purchase negotiation dialogue game  The example deals with buying a new car under some specific constraints such as price, quality, number of seats, etc.  The different communicative acts (also called locutions) in this dialogue game are: