Electronic Institutions

1
Electronic Institutions

• Course Unit 5
• Spezielle Kapitel aus Telekooperation
• Agent-based Cooperation
• Parts are based on Marc Esteva, Juan-Antonio
  Rodriguez-Aguilar, Carles Sierra, Pere Garcia,
  and Josep L. Arcos, On the Formal Modelling of
  Electronic Instiutions, LNAI 1991, 2000

2
Conventions

• Human interactions follow conventions
• Conventions are general agreements on language,
  meaning and behavior
• By following conventions, humans
• Decrease uncertainties about the behavior of
  others
• Reduce conflicts of meaning
• Create expectations about the outcome of the
  interaction
• Simplify the decision process by restricting to a
  limited set the actions that may be taken
• Used in trade, law, games
• On some occasions, conventions become norms
• They establish how interactions must be
  structured within an institution

3
Institutions

• Institutions are the framework within which human
  interaction takes place, defining what
  individuals are forbidden and permitted and under
  what conditions
• Institutions represent the rules of the game
• They not only structure human interactions but
  also enforce individual and social behavior by
  obligating everybody to act according to the
  norms
• Institutions must be created (constitution) and
  may evolve over time (the common law)
• The main characteristic is the clear distinction
  between rules and players
• Examples
• Institution to set and enforce laws (e.g.,
  courts, police) ,
• Institution to monitor and respond to emergencies
  (e.g., ambulance),
• Institution to prevent and recover from disasters
  (e.g., coast guard, fire-fighters),
• etc.

4
Organizations

• Organizations are social units (or human
  groupings) deliberately constructed or
  reconstructed to seek specific goals A.Etz
  ioni (Modern Organizations) 1964
• Organizations include political, economic,
  social, and educational bodies (political
  parties, trade unions, clubs, universities,
  etc.,)
• Main characteristic
• Formalization the attempt to standardize and
  regulate the behavior of the roles interacting
  within an organization
• Human organizations define the roles and
  responsibilities for organizational participants,
  who are expected to bring those into action
  depending on the task and environmental demands
• Roles are standardized patterns of behavior
  required of all players playing a part in a given
  functional relationship in the context of an
  organization
• Organizations must conform to the rules of
  institutions in order to receive legitimacy and
  support

5
What is an electronic institution?

• Electronic institutions are the agents
  counterpart of human organizations
• a framework for enabling, through a communication
  network, interaction between parties according to
  a set of explicit institutional norms and rules
• specifically designed for providing support,
  trust, and legitimacy in agent-based applications
• helps on both providing tools and services for
  and on supervising the intended relationships
  between the parties

6
Major issues

• The organizational approaches do not conveniently
  handle the issues inherent to open MAS
• heterogeneity of agents
• different languages, different objectives
• trust and accountability
• Agents are not expected to follow the rules of
  the society
• exception handling
• detection, prevention and recovery from failures
  that may jeopardize (risking) the global
  operation of the system
• societal change
• capability of accommodating structural changes

7
The FishMarket (Blanes, Girona)
Admission
Exit
Bidding
8
Fish Auction in Blanes
BUYERS ADMISSION
SELLERS ADMISSION
AC
AV
Buyers registration
Fish and sellers registration
AUCTIONEER
S
Fish show and auction
SELLERS SETTLEMENTS
BUYERS SETTLEMENTS
Fishermen payments
Fish delivery and payment
GC
GV
9
Virtual Fish Auction
10
Auction boss
Activates the FishMarket and controls all
auctioning process. It may intervene talking to
other agents. Closes the auction and shuts down
the program. He customizes the program
11
Auction Boss
Sets the auction parameters
12
Auction boss
Controls the auction and closes the it when all
processes are dead.
13
Fish Admission

• The fish admitter interacts with the program
  through a browser which has the following
  functionalities
• Input the fish characteristics for its
  identification and packaging
• classification in boxes
• Initial price setting

14
Auctioneer

• The auctioneer agent interacts with the system
  through a browser in which the actual information
  of the auction is displayed which buyers and
  sellers participate, which round is the auction
  in, what product is being auctioned, initial
  price, etc. The browser offers the following
  functionalities
• Control the auctioning process
• Select the box to auction at any time
• Change the starting price
• Start the round
• Decide on multiple collisions
• Expel buyers due to insufficient credit
• etc.,

15
Auctioneer
16
Buyers
Buyers can interact with the auction house
(buyers admitter, buyers settlements and
auctioneer) through a browser
Buyer identification
Messages to and from the other agents in the
auction house
17
Buyers
Buyers in the auction house
Where in the auction house is the buyer
Buyer credit
Round number
Auctioned product
Seller ID
Seller name
Initial price
Catalogue of the products to be auctioned
18
Buyers
Credit update
Bidding price
Winning price
To bid
19
Buyers
Information in the evolution of the auction
Information on the products sold
To go to different places in the auction house
To bid
To update the credit
To leave the auction house
20
Sellers
Sellers interact with the auction house (sellers
admitter, sellers settlements) through a browser

Seller ID
Messages from the other agents in the market
21
Sellers
Session earnings
Sold products
To leave the market
To include products in the market catalogue
To go to different places in the auction house
22
Electronic Institutions structure

• E-institution consists of three components
• Agents
• encapsulated computer systems that are situated
  in some environment and are capable of flexible,
  autonomous action in that environment in order to
  meet their design goals.
• Interactions
• Agents need to interact with each other in order
  to manage their inter-dependencies.
• These interactions involve agents cooperating,
  negotiating and coordinating with each other.
• Organizations
• The agents' interactions take place within some
  organizational context (e.g. a marketplace or
  some other form of organization)

23
Agents

• Agents interacts with each other by the exchange
  of illocutions
• Each agent has a role (as standardized pattern of
  behavior)
• As dialogic actions are associated to roles, an
  agent adopting a given role is only allowed to
  perform the actions associated with that role
• The identification and regulation of roles is
  considered as part of the formalization process
  of any organization
• In a dialogic organization, the organization
  establishes the acceptable illocutions by
  defining the ontology (vocabulary) - the common
  language to represent the "world"- and the common
  language for communication and knowledge
  representation.
• All of these contextual features are bundled
  together in a dialogic framework

24
Interaction

• Interactions between agents are articulated
  through agent group meetings, which is called
  scenes, with a well-defined communication
  protocol (the possible dialogues agents may have)
• Scenes can be connected, composing a network of
  scenes (the so-called performative structure)
  which captures the existing relationships among
  scenes.
• The specification of a peformative structure
  contains a description of how different
  agents/roles can legally move from one scene to
  another.
• A performative structure is to contain the
  multiple, simultaneous ongoing activities,
  represented by scenes
• Agents within a performative structure may
  participate in different scenes at the same time
  with different roles.

25
The fundamental elements

• Dialogic Framework
• Linguistic and ontological conventions to make
  efficient communication among agents.
• Performative Structure
• Activities specified as connections among scenes.
• Norms
• Constraints on scenes

26
Dialogical Framework

• We define a dialogical framework as a tuple
• DF ltO, I, L, RI, RE, RSgt
• where
• O stands for the ontology
• I is the set of illocutionary particles
• L stands for a representation language
• RI is the set of internal roles
• RE is the set of external roles and
• RS is the list of relationships over roles

27
Elements of a Dialogic Framework
dialogic-framework a name with which to refer to
this framework. ontology a name referring to a
defined ontology content-language a name
defining the content language (e.g., KIF, PROLOG
or LISP). illocutionary-particles a list of
names of illocutionary particles to be used in
the illocutions. external-roles a list of names
of roles that external agents may
play. internal-roles a list of names of roles
that internal (staff) agents may
play. social-structure a list of triples of two
role names and the name of relationship between
them.
28
The FishMarket and Buyer Settlement Scene
Dialogic Framework
(define-dialogic-framework fm-dialogic-framework
as ontology fm-ontology content-language
PROLOG illocutionary-particles (request accept
deny inform commit pay) external-roles (buyer
seller) internal-roles (boss buyer-admitter
seller-admitter auctioneer buyer-accountant
seller-accountant) social-structure ((boss lt
buyer-admitter) (boss lt seller-admitter)(boss lt
auctioneer) (boss lt buyer-accountant) (boss lt
seller-accountant) (buyer incompatible
seller)) ) (define-dialogic-framework buyer-settl
ement-df as ontology buyer-settlement-ontology c
ontent-language PROLOG illocutionary-particles
(request inform accept deny pay) )
29
FishMarket Dialogical Framework
30
Scenes

• Scene is a pattern of multi-agent conversation.
• Scene is specified by a finite state oriented
  graph where the nodes represent the different
  states and oriented arcs are labelled with
  illocution schemes or timeouts.
• During the execution new agents can join the
  scene or some of the participants can leave the
  scene at definite states depending on their role.
• An scene can be multiple-instantiated and played
  by different groups of agents.

31
Elements of a Scene Specification
roles a list of names of role that may
participate in the scene. dialogic-framework the
name of the dialogic framework to be used for
communication within the scene. states a list of
the names of the states of the conversation
graph. initial-state a name identifying the
initial state. final-states a list of names
identifying final states. access-states a list
of pairs of role name and a list of states,
identifying which roles may join at which
states. exit-states a list of pairs of role name
and a list of states, identifying which roles may
leave at which states. agents-per-role a list of
triples of role name, minimum integer and maximum
integer, defining the constraints on the
population of a particular role. connections a
list of the transitions between scene states.
Each one comprises a preceding state name, a
succeeding state name, and either an
illocution-scheme with some constraints over
scenes variables which must be satisfied to
progress through this transition or a timeout
that will trigger the transition when will expire.
32
FishMarket Buyer Settlement Scene
(define-scene buyer-settlement-scene as roles
(buyer buyer-accountant) scene-dialogic-framework
buyer-settlement-df states (w0 w1 w2 w3 w4 w5
w6 w7) initial-state w0 final-states
(w7) access-states ((buyer (w0 w2 w3
w6)) (buyer-accountant (w0)) ) exit-states
((buyer (w2 w3 w6 w7)) (buyer-accountant
(w7))) agents-per-role ((0 lt buyer lt 1) (1 lt
buyer-accountant lt 1) ) connections ((w0 w1
buyer-settlement-i1) (w1 w2 buyer-settlement-i2) (
w1 w3 buyer-settlement-i3) (w0 w4
buyer-settlement-i4) (w4 w5 buyer-settlement-i5) (
w5 w6 buyer-settlement-i6) (w2 w1
buyer-settlement-i1) (w3 w1 buyer-settlement-i1) (
w6 w1 buyer-settlement-i1) (w2 w4
buyer-settlement-i4) (w3 w4 buyer-settlement-i4) (
w6 w4 buyer-settlement-i4) (w2 w7 closing) (w3 w7
closing) (w6 w7 closing) ) )
33
Auction Scene
34
Performative Structure

• Complex activities can be specified by
  establishing relationships among scenes that
• capture causal dependencies.
• define synchronisation mechanisms.
• establish parallelism mechanisms.
• define choice points that allow roles leaving a
  scene to choose which activity to engage in next.
• establish the role flow policy.

35
Performative structure

• A performative structure can be seen as a network
  of scenes.
• We introduce transitions to mediate between
  connections of scenes.
• Arcs connecting scenes and transitions labelled
  with constraints.
• The specification allows to express that an scene
  can be running simultaneously multiple-times at
  execution time.
• Determines whether agents moving between scenes
  join current executions of the target scene(s) or
  whether new executions are started.

36
Elements of the Performative Structure

• scenes a list comprising a name for the scene,
  the class of the scene
• transitions a list comprising a name for the
  transition and the class of the transition.
• connections a list containing the connections
  from scenes to transitions and from transitions
  to scenes. In the first case the connection is
  expressed by the source scene name, the target
  transition name, a list of pairs of
  agent-variable and role name, and a list of
  constraints that will restrict agents movements.
  In the second case is expressed by the source
  transition name, the target scene name, a list of
  pairs of agent-variable and role name, and a name
  defining if a new execution of the scene will be
  created or if the agent(s) will go to one, some
  or all current executions.
• initial-scene the name of the initial scenefrom
  one of those given in scenes.
• final-scene the name of the final scenefrom one
  of those given in scenes.

37
Fishmarket Performative Structure

• (define-performative-structure
• fm-performative-structure as
• scenes ((enter root-scene) (exit output-scene)
• (buyer-admission buyer-admission-scene)
• (seller-admission seller-admission-scene)
• (auction-room auction-room-scene list)
• (buyer-settlement buyer-settlement-scene)
• (seller-settlement seller-settlement-scene))
• transitions ((t1 AND-AND) (t2 AND-AND) (t3
  AND-AND)
• (t4 AND-AND) (t5 AND-AND) (t6 AND-AND)
• (t7 AND-AND) (t8 AND-AND) (t9 AND-AND)
• (t10 AND-AND) (t11 AND-AND) (t12 AND-AND) (t13
  AND-AND) )
• connections
• ((enter t1 ((x buyer-admitter)))
• (t1 buyer-admission ((x buyer-admitter)) new)
• (enter t2 ((x buyer)))
• (t2 buyer-admission ((x buyer)) one)
• (enter t3 ((x seller-admitter)))
• (t3 seller-admission ((x seller-admitter)) new)

38
FishMarket Performative Structure
39
Norms

• Norms are mechanisms that a society has in order
  to influence the behaviour of the agents
• Norms can be created from different sources,
  varying from built-in norms to simple agreements
  between agents, or more complex legal systems
• Characteristics
• Norms are always prescribed to be compiled with
  for a set of agents i.e., they specify something
  that ought to be done and consequently they
  include normative goals that must be addressed
• Norms are always applicable, and their activation
  depends on the context in which agents are
  situated i.e., there maybe exceptions where
  agents are not obliged to comply with the norm
• Norms suggest the existence of a set of sanctions
  or punishments to be imposed when an agent
  doesnt satisfy the normative goal, and a set of
  rewards to be received when agents do.

40
Examples

• Norms
• Everybody arrives at 900 (statistically)
• This can be used to predict the presence of
  people
• The norm is that everybody arrives at 900
• The presence of this norm is a factor for each
  person to decide the time to come to work
• Obligations
• Whenever your boss assigns you a task you have to
  fulfil the task as quick as possible
• If you dont comply, your boss can reprimand you
  and perhaps fire you

41
Norms

• The normative rules defines the consequences of
  agents actions within the institution.
• Such consequences are
• some actions can impose obligations to agents.
• can vary the paths that agents can follow.
• Norms are specified by three elements
• Antecedent the actions that provoke the
  activation of the norm and restrictions over
  illocution scheme variables.
• Defeasible antecedent the actions that agents
  must carry out in order to fulfill the
  obligations.
• Consequent the set of obligations
• For instance, a buyer winning a bidding round is
  obliged to go later on to the buyers settlement
  scene to pay for the good.

42
Norm Specification

• antecedent a list comprising an arbitrary number
  of pairs of scene name and illocution scheme name
  and a list of boolean expressions over illocution
  scheme variables.
• defeasible-antecedent a list comprising an
  arbitrary number of pairs of scene name and
  illocution-scheme name.
• consequent a list of obl predicates.

43
The Obligation2pay and Sanction norms

• (define-norm obligation2pay as
• antecedent
• ((auction-room
• (inform (?y auctioneer) (?x buyer) (sold(?good-id
  ?price ?x)))))
• defeasible-antecedent
• ((buyer-settlement (inform (!x buyer) (?y
  buyer-accountant)
• (payment(!price)))))
• consequent
• ((obl !x (inform (!x buyer) (?y buyer-accountant)
• (payment(!price))) buyer-settlement))
• )
• (define-norm sanction as
• antecedent
• ((auction-room
• (commit (?x buyer) (?y auctioneer) (bid(?good-id
  ?price))))
• (lt (credit !x) !price))
• defeasible
• ((auction-room (inform (!y auctioneer) buyer
  (sanction(!x)))))
• consequent

44
Cooperation vs. Coordination

• Two approaches for designing and modelling norms
  in electronic institutions
• Cooperative normative behaviour
• Norms are defined implicitly in the agent
  architecture and depends on the how agents
  function in an institution
• This is based on the assumption that agents have
  common goals and intentions which enforce some
  kind of cooperation
• Coordination strategy
• Norms are defined explicitly in the institution
  and agents have to comply to them
• Norms are nothing but a an instrument of
  individual, collective, or generic optimization

45
Norms Abstractions

• Values
• Values are the embodiment of what an organization
  stands for, and should be the basis for the
  behavior of its members.
• Examples are those value related to the integrity
  of the institution, professionalism and
  excellence, and contractual obligations
• Normative Behavioural Rules
• These are the rules of conduct within the
  institution.
• They are necessary to judge if the behavior of
  the members of the institution complies with the
  values defined on the first level of the
  abstraction
• Examples are declaration of interest
  (registration), accuracy of information, and fair
  competition.
• Norms
• This is the lowest level of abstraction
• Examples are social norms (organization, roles,
  authority), communication norms (protocols,
  dialogues), and behavioral norms (obligations,
  permissions, interdiction)

46
Hierarchical Meta Model of the Normative Structure

• Values
• logical facts about the institution
• institution(fair, responsible)
• agent(polite, cooperative, helpful)
• information(accurate)
• Normative Rules
• clausal formulas
• transaction(secure)?institution(fair, -),
  agent(polite,-,-) ,information(accurate)
• Norms
• are restricted to denial constraints that is,
  formulas of the form
• ? L1 ,, Ln
• there is no question asked and dont get answer
  in a multi-agent system
• ? agent(polite,-,-), ask(ai , aj, ?), ans(aj ,
  ai, ?)
• there is no age discrimination in an electronic
  institution
• ? institution (fair,-), age(young), age(old)
• there is no entrance to the institution without
  registration
• ? institution(-,responsible)

47
Example I
48
Example II
49
Example Scenario

• Suppose we wish to model the following norms
  through an institution
• before participating in an auction, buyers and
  sellers are obliged to register.
• if bid accepted in an auction, buyer is obliged
  to pay
• if payment received in settlement, seller is
  obliged to deliver good
• if has no obligation, buyer or seller may leave.
• scenes
• registration of buyers and sellers
• auction scene (where a standard downward
  bidding/Dutch auction format is employed)
• the settlement scene where payments and delivery
  of goods can take place.

50
Performative Structure
51
Electronic Institution

• An electronic institution is defined as a tuple
  EI ltDF, PS, Ngt where
• DF stands for a dialogic framework.
• PS stands for a performative structure.
• N stands for a set of norms.

52
Pros

• Agents working under norms do not need to
  calculate continuously their utilities and,
  therefore, do not need complete information
• Agents are supposed to act in a somehow
  predetermined way according to the principle of
  mutual expectation
• Norms imply that the agents respect certain
  social constraints that deter them from breaking
  agreements

53
Cons

• Agents are assumed to follow rules just because
  they are designed to do so
• Agents are not seen as autonomous entities any
  more!
• Proposals so formulated are therefore closer to
  Distributed AI than to MAS
• Solution?

54
(No Transcript)