Transcript Electronic Institutions: from specification to development

IDE-eli: An Integrated Development
Environment for Electronic Institutions
Carles Sierra, Bruno Rosell, Juan A. Rodríguez-Aguilar, Pablo Noriega,
Marc Esteva, Guifré Cuní, David de la Cruz, Josep Ll. Arcos
Artificial Intelligence Research Institute, IIIA
Spanish Scientific Research Council, CSIC
Introduction
• Open
•
•
multi-agent systems are populated by
heterogeneous and self-interested agents, developed by
different people, using different languages and
architectures.
Research issue: need for appropriate methodologies and
software tools which give support to the analysis, design,
and development of open systems.
Goal: principled design and development of open multi
agent systems.
2
Approach
Institutions in the sense proposed by North “… set of artificial
constraints that articulate agent interactions”.
3
EI Development cycle
4
Methodology and software tools in IDE-eli
SPECIFICATION
e-INSTITUT ION
SPECIFICATION
VIA ISLANDER
IDE-eli
ISLANDER
VERIFICATION
VERIFY STATIC
PROPERTIES
[NO]
SOUND?
aBuilder
[YES]
AGENT
SYNTHESIS &
CUSTOMISATION
DEFINITION OF
PROTOTYPE
SimDei
SIMULAT ION
ANALYSE
RESULTS
[NO]
[YES]
AMELI
[NO VALID]
CHANGE
SPECIFICATION?
[VALID]
[YES]
AGENT
BUILDING
CODE GENERATION
CHANGE
SPECIFICATION?
[NO]
[NOT OK]
ELECT RONIC
INSTIT UTION
EXECUTION
ELECT RONIC
INSTIT UTION
MONIT ORING
EXECUTION
ANALYSIS
5
EXECUTION &
ANALYSE
Monitoring Tool
Electronic Institution Specification with ISLANDER
 Common Ontology and
language
 Agent Roles
 Multi-agent Protocols
 Network of protocols
 Norms
6
Electronic Institution Specification with ISLANDER
PERFORMATIVE STRUCTURE
(NETWORK OF PROTOCOLS)
SCENE
(MULTI-AGENT PROTOCOL)
ROLES
NORMS
Buyers’ Payment
7
Electronic Institution simulation with SIMDEI
eI.xml
SIMDEI
Observables
Expected Behavior
eI model
RePast
8
analysis
Electronic Institution simulation with SIMDEI
9
Agent development with aBuilder
10
Electronic Institution Infrastructure
Our approach
Usual approach
E.Institution
Autonomous
Agents Layer
Autonomous
Agents Layer
Agent 1 . . .
Agent 1 . . .
Agent n
Agent n
AMELI
JADE
JADE
Communication
Layer
Communication
Layer
11
AMELI architecture
Ai
...
An
G1
...
Gi
...
Gn
IM
S M1
...
...
S Mm
T M1
... T Mk
-
-
Communication Layer
12
Participating
Agents Layer
Public
...
Private
Institution
Specif ication
(XML
f ormat)
A1
Social layer
(AMELI)
AMELI functionalities
• MEDIATION
•
• To facilitate interagent communication within conversations.
COORDINATION AND ENFORCEMENT
• To guarantee the correct evolution of each conversation
(preventing errors made by the participating agents by filtering
erroneous illocutions, thus protecting the institution).
• To guarantee that agents’ movements between scenes comply with
the specification.
• To control which obligations participating agents acquire and
fulfil.
• INFORMATION MANAGEMENT
• To facilitate participating agents the information they need to
successfully participate in the institution.
13
Simulating electronic institutions
OBJECT LEVEL
AGENT LEVEL
The Model
ELECTRONIC
INSTITUTION
(environment)
Schedule
Agents
Probes
The Interface
14
Test and Deployment with AMELI - Monitoring
15
Test and Deployment with AMELI - Monitoring
16
AMELI implementation features
•
•
•
•
•
•
Agent-based
Realised as a middleware layer
Architecturally neutral
General purpose (can interpret any institution specification)
Communication neutral
Scalable (it can be distributed among several machines)
17
Conclusions
• Engineering open multi-agent systems is a highly complex
•
•
task.
Electronic institutions introduce normative (regulatory)
environments that structure agent interactions, establishing
at each moment what agents are permitted and forbidden to
do.
We have presented an electronic institutions development
environment that supports the specification, verification,
analysis, execution, and monitoring of open multi-agent
systems.
18