MultiAgent Systems Lecture 6 University Politehnica of Bucarest 2004 2005 Adina Magda Florea adinacs

1
Multi-Agent SystemsLecture 6 University
Politehnica of Bucarest2004 - 2005Adina
Magda Floreaadina_at_cs.pub.rohttp//turing.cs.pub
.ro/blia_2005
2
Coordination mechanisms and strategiesLecture
outline

• 1 Coordination strategies
• 2 Modeling coordination through AND/OR graphs
• 3 Modeling coordination by shared mental states
• 4 Joint action and conventions

3
1 Coordination strategies

• Coordination the process by which an agent
  reasons about its local actions and the
  (anticipated) actions of others to try to ensure
  the community acts in a coherent manner

Coordination
Self-interested agents own goals
Collectively motivated agents common goals
Coordination for coherent behavior
Cooperation to achieve common goal
Competitive conflicting goals
Neutral to one another disjunctive goals
3
4

• Model
• Protocol
• Communication

• Perfect coordination ???
• Centralized coordination ?
• Distributed coordination

• Tightly coupled interactions - distributed search
• Complex agents - distributed planning, task
  sharing, resource sharing
• Heterogeneous agents - interaction protocols
  Contract Net, KQML conversations, FIPA protocols
• Dynamic interactions - Meta-level information
  exchange, commitments and conventions
• Complex interactions - organizational structure
  to reduce complexity
• Unpredictable interactions - social laws
• Conflict of interests - interaction protocols
  voting, auctions, bargaining, market mechanisms,
  extended Contract Net, coalition formation

Cooperative
Neutral or competitive
4
5
2 Modeling coordination through AND/OR graphs

• Activities of the agents represented as a search
  through an AND/OR goal graph
• AND/OR goal graph augmented with a representation
  of dependencies
• between goals
• primitive goals and resources needed to solve
  them
• Interdependencies
• weak or strong
• uni-directional or bi-directional
• Joint goals - a team of agents decide to pursue a
  common goal in a cooperative manner
• Joint goals must be mapped into individual goals

5
6
AND/OR goal graph with dependencies between
goals and shared resources
Agent2
Agent1
Find the most consistent explanation of sensory
data
G10
G20
OR
AND
G1,2m
.
G2t
G2p
.
G11
G1k
G12
Find vehicle tracks in a narrowly defined region
AND
AND
OR
G1m,1
G11,2
G11,1
G2p,2
G2p,1
G2m,2
OR
AND
Identify the types of vehicle present based on
sensory data
AND
G2p,2,2
G1m,1,1
G2p,1,4
G1m,1,2
G2p,1,3
d2z
DATA/ Resources
d11
d1j
d2j1
6
7
3 Modeling coordination by shared mental states

• Based on the view of intentional stance agents
• Example of intentional coordinated action
• 3.1 Collective mental states
• (a) Common knowledge
• Every member in group G knows p EGp ??ai?GKaip
• - shared knowledge
• Every member in G knows EGp E2Gp ? EG(EGp)
• Every member knows that every member knows that
  every
• Ek1Gp ? EG(EKGp) k?1
• Common knowledge
• CGp ? p ? EGp ? E2Gp ? ? EkGp ? ...

7
8

• (b) Mutual belief
• EGp ??ai?GaiBelp - Every one in group G believes
  p - shared belief
• E2Gp ? EG(EGp)
• Ek1Gp ? EG(EKGp) k?1
• MGp ? EGp ? E2Gp ? ? EkGp ? - Mutual belief
• ? Perfectly shared mental state but mutual
  beliefs (as common knowledge) can not be
  guaranteed because communication between agents
  is not reliable in terms of delivery and delay
• (c) Joint intentions (Levesque Cohen, 1990)
• C1) each agent in the group has a goal p
• ?ai?G aiIntp (and cf goal-intentions
  compatibility aiIntp ? aiDesp)
• C2) each agent will persist with this goal until
  it is mutually believed that p has been achieved
  or that p cannot be achieved
• ?ai?G aiInt (A Fp) ? A ( aiInt(A Fp) ?
• (MG(Achieve p) ? MG(?Achieve p)))
• C3) conditions (C1) and (C2) are mutually
  believed
• MG(C1) ? MG(C2)

8
9

• Commitments
• Formal
• ? Blindly committed agent
• xInt(A Fp) ?A (xInt(A Fp) ? xBelp)
• ? Single-minded committed agent
• xInt(A Fp) ?A (xInt(A Fp) ? (xBelp ? ?xBel(E
  Fp)))
• ? Open-minded committed agent
• xInt(A Fp) ?A (xInt(A Fp) ? (xBelp ? ?xDes(E
  Fp)))
• Informal Commitments may be seen as pledges
  about beliefs and actions
• (d) Joint commitments
• Agents in the group the state of joint
  commitment is distributed
• ? have a joint goal the group becomes
  jointly committed
• ? agree they wish to cooperate to achieve
  the goal (joint goal)
• Joint intentions can be seen as a joint
  commitment to a joint action while in a certain
  shared mental state

F - eventually G - always A - inevitable E -
optional
9
10

• 4 Joint action and conventions
• 4.1 Conventions
• An agent should honor its commitments provided
  the circumstances do not change.
• Conventions describe circumstances under which
  an agent should reconsider its commitments
• An agent may have several conventions but each
  commitment is tracked using one convention

10
11

• Commitments provide a degree of predictability so
  that the agents can take future activity of other
  agents in consideration when dealing with
  inter-agent dependencies ? the necessary
  structure for predictable interactions
• Conventions constrain the conditions under which
  commitments should be reassesed and specify the
  associated actions that should be undertaken
  retain, rectify or abandon the commitment ? the
  necessary degree of mutual support

11
12

• 4.2 Specifying conventions
• Reasons for re-assessing the commitment
• commitment satisfied
• commitment unattainable
• motivation for commitment no longer present
• Actions
• R1 if commitment satisfied or
• commitment unattainable or
• motivation for commitment no
  longer present
• then drop commitment
• But such conventions are asocial constructs they
  do not specify how the agent should behave
  towards the other agents if
• it has a goal that is inter-dependent
• it has a joint commitment to a joint goal

12
13

• Social Conventions
• Invoke when Inter-dependent goals
• local commitment dropped
• local commitment satisfied
• motivation for local commitment no longer present
• R1 if local commitment satisfied or
• local commitment dropped or
• motivation for local commitment no
  longer present
• then inform all related commitments
• Invoke when Joint commitment to a joint
  goal
• status of commitment to joint goal changes
• status of commitment to attaining joint goal in
  the team context changes
• status of commitment of another team member
  changes
• R1 if status of commitment to joint goal
  changes or
• status of commitment in the team
  context changes
• then inform all other team members of
  the change
• R2 if status of commitment of another team
  member changes
• then determine whether joint
  commitment is still viable

13
14

• 4.3 An example of joint action and conventions
• GRATE System (Generic Rules and Agent model
  Testbed Environment, Jennings, 1994)
• ARCHON electricity distribution management
• cement factory control
• Electricity distribution management of the
  traffic network
• distinguish between disturbances and pre-planned
  maintenance operations
• identify the type (transient or permanent),
  origin and extend of faults when they occur
• determine how to restore the network after a
  fault
• 3 agents
• AAA - the Alarm Analysis Agent ? perform
  diagnosis to different levels
• BAI - the Blackout Area Identifier of precision
  and on different info
• CSI - Control System Interface ? detects the
  disturbance initially and then monitors
  the network evolving state

14
15
GRATE Agent Architecture
Inter-agent communication
Communication Manager
CONTROL DATA
COOPERATION MODULE
Cooperation Control Layer
SITUATION ASSESMENT MODULE
CONTROL MODULE
Domain Level System
Task3
Task1
Task2
15
16

• (a) Agent behavior
• 1. Select goal and develop plan to achieve goal
• 2. Determine if plan can be executed individually
  or cooperatively
• (a) joint action is needed (joint goal) or
• (b) action solved entirely locally
• 3. if (a) then the agent becomes the organiser
• 3.1. Establish joint action - the organiser
  carries on the distributed planning protocol
• 3.2. Perform individual actions in joint action
• 3.3. Monitor joint action
• 4. if (b) then perform individual actions
• 5. if request for joint action then the agent
  becomes team-member
• 5.1. Participate in the planning protocol to
  establish joint action
• 5.2. Perform individual actions in joint actions
• (3.2 and 5.2 adequately sequenced)

16
17
(No Transcript)
18

• Joint intention - Phase 1 for agent AAA
• Name Diagnose-fault
• Motivation Disturbance-detection-message
• Plan S1 Identify_blackout_area, S2
  Hypothesis_generation,
• S3 Monitor_disturbance, S4
  Detailed_diagnosis, S2 lt S4
• Start time Maximum end time
• Duration Priority 20
• Status Establish group
• Outcome Validated_fault_hypothesis
• Participants ((Self organiser agreed_objective)
• (CSI team-member agreed_objective)
• (BAI team-member agreed_objective))
• Bindings NIL
• Proposed contribution
• ((Self (Hypothesis_generation yes)
  (Detailed_diagnosis yes))
• (CSI (Monitor_disturbance yes)
• (BAI (Identify_blackout_area yes)))

18
19

• Joint action - Phase 2 for agent AAA
• Name Diagnose-fault
• Motivation Disturbance-detection-message
• Status Establish plan
• Start time 19
• Maximum end time 45
• Bindings ((BAI Identify_blackout_area 19 34)
• (Self Hypothesis_generation 19 30)
• (CSI Monitor_disturbance 19 36)
• (Self Detailed_diagnosis 36 45))
• .
• BAI's individual intention for producing the
  blackout area
• Name Achieve Identify_blackout_area
• Motivation Satisfy Joint Action Diagnose-fault
• Start time 19 Maximum end time 34
• Duration 15 Priority 5
• Status Pending
• Outcome Blackout_area

19
20

• (c) Monitor the execution of joint action
• Recognize situations that change commitments and
  impact joint action
• R1match if task t has finished executing and
• t has produced the desired outcome of the
  joint action
• then the joint goal is satisfied
• R2match if receive information i and
• i is relevant to the triggering
  conditions for joint goal G and
• i invalidates beliefs for
  wanting G
• then the motivation for G is no
  longer present
• Social conventions
• R1inform if joint action has successfully
  finished
• then inform all team members of
  successful completion and
• see if result should be
  disseminated outside the team
• R2inform if motivation for joint goal G is no
  longer present
• then inform other members of the
  team that G needs to be abandoned
• Rules to indicate what to do if change in
  commitments
• ..

20
21

• References
• Multiagent Systems - A Modern Approach to
  Distributed Artificial Intelligence, G. Weiss
  (Ed.), The MIT Press, 2001, Ch.2.3, 8.5-8.7
• V.R. Lesser. A retrospective view of FA/C
  distributed problem solving. IEEE Trans. On
  Systems, Man, and Cybernetics, 21(6), Nov/Dec
  1991, p.1347-1362.
• N.R. Jennings. Coordination techniques for
  distributed artificial intelligence. In
  Foundations of Distributed Artificial
  Intelligence, G. O'hara, N.R Jennings (Eds.),
  John WileySons, 1996.
• N.R. Jennings. Controlling cooperative problem
  solving in industrial multi-agent systems using
  joint intentions. Artificial Intelligence 72(2),
  1995.
• E.H. Durfee. Scaling up agent coordination
  strategies. IEEE Computer, 34(7), July 2001,
  p.39-46.

21