testbed in the Do we need a testbed in the Do we need a COIN - - PowerPoint PPT Presentation

testbed in the do we need a testbed in the do we need a
SMART_READER_LITE
LIVE PREVIEW

testbed in the Do we need a testbed in the Do we need a COIN - - PowerPoint PPT Presentation

testbed in the Do we need a testbed in the Do we need a COIN community and for what ? COIN community and for what ? COIN @ AAMAS Held with AAMAS May 12, 2008, Estoril, Portugal Examples of Testbeds and Contest Examples of Testbeds and


slide-1
SLIDE 1

Do we need a Do we need a testbed testbed in the in the COIN community and for what ? COIN community and for what ?

COIN @ AAMAS

Held with AAMAS May 12, 2008, Estoril, Portugal

slide-2
SLIDE 2

Examples of Testbeds and Contest Examples of Testbeds and Contest

Artificial Intelligence

 Turing Test  RoboCup

Agent Technology

 Trading Agent Competition  Agent Contest (ProMAS)

Trust and Reputation

 ART Testbed

slide-3
SLIDE 3

Objectives Objectives

Motivation & Challenge

 to find out solutions for a complex problem

Comparison & evaluation

 qualitative, quantitative  models, languages, platforms

Improvement of the quality of tools in the domain

slide-4
SLIDE 4

What to compare? What to compare?

Languages and Models of

 Coordination  Organisations  Institutions  Norms

Platforms of COIN

 results from platforms can be extended to the underlying model (?)

 depends of how to compare

slide-5
SLIDE 5

Which metrics? Which metrics?

 expressiveness  evolution & adaptation  flexibility  reactiveness  requirements for the agent's cognitive capabilities  constraints x autonomy  scalability  reliability  maturity

 #case studies

 cost

 #messages

 ...

slide-6
SLIDE 6

How to compare? How to compare?

by developing Testbed

 Dist. Vehicle Monitoring Testbed, ART Testbed

by organising Contest

 Agent Contest, ART Contest

Based on the performance in a scenario Avoid the Turing Test problem (?)

 Considering just the “result” (e.g. ART Contest)  Considering both the “result” and the “method” (e.g. Agent Contest)

slide-7
SLIDE 7

Scenario – assumptions Scenario – assumptions

Open

 agents can enter and leave the systems

Any agent architecture

 no access to the “internals” of the agents

slide-8
SLIDE 8

Scenario – desired properties Scenario – desired properties

Cooperation

 no single agent can have a good result

Explicit rules must be available

 to allow agents to reason about them  even violating them

Require different dimensions

 structural, functional, dialogical, normative, ...

slide-9
SLIDE 9

Task accomplishment scenario (TAS) Task accomplishment scenario (TAS)

 The platform should help the collective accomplishment of some task  Purposeful

 describe goals, activities, objectives, etc

 Measurable

 objective ways to evaluate the performance of the platform

 Feasible

 a task that may be accomplished by software agents

slide-10
SLIDE 10

Metrics for TAS Metrics for TAS

 time to accomplish  allocation of goals to capable agents  cost of coordination  constraints  detection of violations  sanctions  e.g. remove bad agents from the system  scalability  e.g. size and number of tasks  ...

slide-11
SLIDE 11

agent

Method I Method I

given by the contest given by the participant

agent agent ....... agent

Platform

given by the contest

agent agent agent agent

Evaluator

slide-12
SLIDE 12

Method II Method II

given by the participant given by the contest

agent agent agent

Platform

agent

Evaluator

slide-13
SLIDE 13

agent

Method III Method III

given by the participant given by the contest

agent agent ....... agent

Platform

given by the participant

agent agent agent agent

Evaluator

slide-14
SLIDE 14

Discussion Discussion

What ? Language, Model, Platform Metrics? Testbed or Contest? Which scenario? TAS, TAC, ... Which Method (I, II, III)? Suggestions, New ideas?