CPSC%20601.68/CPSC%20599.68

1
Agent Communications

• CPSC 601.68/CPSC 599.68
• Rob Kremer
• Department of Computer Science
• University of Calgary

2
What is an Agent?

• Autonomous
• Knowledge-based
• Reflective or Reactive
• Social

3
What is an Agent?

• An agent Ag is a tuple Ag (Sit,Act,Dat,fAg)
  where
• Sit is a set of situations Ag can be in
• Act is a set of actions Ag can perform
• Dat, Ags internal data, is the set of possible
  values Ags internal data areas can have
• fAg Sit ?Dat ? Act is the agents decision
  function
• From outside, Ag seems to be a function
  gAgSit?Act

There are lots of other models of agency this
one is usefully general
4
Agents Object?

• Both have internal data and receive messages
• But the object has no choice but to execute the
  method the agent can decide if and what method
  to execute.

I might execute method B, or I might do nothing,
or I might execute method C or .
object
agent
message B
sensor
method A
message B
method B
method A
method C
method B
method C
5
Communicating Agents

• Traditional (low level) Protocols (eg TCP/IP)
  are a bit simple
• They mostly deal with heres a message
• Agents need a richer conversation space
• Request/Reply, Offer/Reply, Subscribe/Inform
• Agents need to handle exceptional cases, complex
  interactions, and novel new interactions

6
Simple Conversation naive
What time is it?
Its 200
7
Simple Conversation more complex
What time is it?
Just a sec, Ill check
Its 200
Thanks
8
Simple Conversation failure
What time is it?
I dont know, my watch stopped
9
What do we need?

• Message format (syntax)
• Message semantics
• Interpretation of sequences of messages
  (conversations)
• Message sequencing and turn-taking
• Conversation semantics

10
Message Syntax

• Envelope/content
• Envelope can be read by all agents in the society
• Content is usually understood only by specialized
  agents

11
Message Syntax
performative the type of the communicative act of the message
sender the sender of the message
receiver the intended recipients of the message
reply-to subsequent messages in this conversation thread are to be directed to the agent named
content the content of the message equivalently denotes the object of the action
language the language in which the content parameter is expressed
encoding the specific encoding of the content field
ontology the ontology(s) used to give a meaning to the symbols in the content expression
protocol the interaction protocol that the sending agent is employing with this ACL message
conversation-id an expression used to identify the sequence of communicative acts that form a conversation
reply-with an expression that will be used by the responding agent to identify this message
in-reply-to an expression that references an earlier action to which this message is a reply
reply-by a time and/or date expression indicating the latest time the sending agent wants a reply
12
Message Syntax KQML-style

• ( request
• act inviteToJoinCD
• to casa//123.181.6.101/Bob
• from casa//123.181.6.101/Alice
• reply-with casa//123.181.6.101/Alice1453
• language casa.URL
• language-version 1.0
• content casa//123.181.6.101/CDagent2
• )

13
Message Syntax xml-style

• lt?xml version\"1.0\"?gt
• lt!DOCTYPE message SYSTEM \"xmlmessage.dtd\"gt
• ltCASAmessagegt
• ltperformativegt request lt/performativegt
• ltactgt inviteToJoinCD lt/actgt
• lttogt casa//123.181.6.101/Bob lt/togt
• ltfromgt casa//123.181.6.101/Alice lt/fromgt
• ltreply-withgt casa//123.181.6.101/Alice1453
  lt/reply-withgt
• ltlanguagegt casa.URL lt/languagegt
• ltlanguage-versiongt 1.0 lt/language-versiongt
• ltcontentgt casa//123.181.6.101/CDagent2
  lt/contentgt
• lt/CASAmessagegt
• lt/xmlgt

14
Message Semantics

• Typically based on speech-act theory
• Performative is the main speech act
• FIPA implementation uses the performative and a
  BDI semantics

15
Message Semantics FIPA Performatives
accept-proposal accepting a previous proposal
agree agreeing to perform some action
cancel inform another agent that the agent no longer need perform some action
call-for-proposal calling for proposals to perform an action
confirm informs a given proposition is true
disconfirm informs a given proposition is false
failure an action was attempted but failed
Inform a given proposition is true
inform-if inform whether or not a proposition is true
inform-ref inform the object which corresponds to a descriptor
not-understood did not understand what the receiver just did
propagate pass a message on
propose submitting a proposal to perform an action
proxy pass on an embedded message
query-if asking whether or not a proposition is true
query-ref asking for the object referred to
refuse refusing to perform an action
reject-proposal rejecting a proposal during negotiation
request request to perform some action
request-when request to perform some action when some proposition becomes true
request-whenever request to perform some action each time the proposition becomes true
subscribe requesting to notify of the value of a reference whenever the object changes
16
Message Semantics FIPA request
lti, request (j, a)gt FP FP (a) i\j ? Bi Agent
(j, a) ? Bi Ij Done (a) RE Done (a) FP(a)
i\j denotes the part of the FPs of a which are
mental attitudes of i.
17
Message Semantics Social Commitments/CASA
18
Conversation Paradigms

• Ad-hoc protocols
• BDI (Belief, Desire, Intentions)
• Social Commitments

19
Ad-hoc Protocols

• Scripts that the agent follows
• Issues
• Rigid and Brittle.
• Hard to account for all eventualities

20
Ad-hoc Protocols Petri Nets
21
BDI

• Belief
• Environment
• Desire
• Goals
• Intention
• The current desire(s) chosen by the selection
  function with the beliefs and desires as input

22
BDI

• Issues
• Sometimes end up with expressions like I believe
  that you believe that I believe that you believe
  that
• Calls on agents to have a omniscient view of all
  the other agents
• FIPA is based on BDI semantics

23
Social Commitments

• Many utterances imply some sort of
  conversational social commitment
• Eg a request commits the receiver to reply
• Other social commitments are negotiated
• Eg wash my car
• Basic Agent body
• When an agent observes (or sends or receives) a
  message, it uses policies (rules) as social norms
  that generate (or delete) social commitments
• Agents spend their free time trying to fulfill
  the social commitments for which it is a debtor

24
Social Commitments
Alice
Bob
25
Social Commitments

• Issues
• Deals with complexity
• Agents make decisions based only on observable
  conversational moves (and the Environment and
  their internal state)
• Turn taking follows naturally