RE: [jade-develop] Automatic generation of Finite State Machines


Subject: RE: [jade-develop] Automatic generation of Finite State Machines
From: Robert Kessler (kessler@cs.utah.edu)
Date: Thu Oct 10 2002 - 17:43:09 MET DST


Dick Cowan, Martin Griss, Steve Fonseca and I have done a fair bit of
work on putting UML State charts into Jade. We have several papers that
we wrote about it. The relevant ones are:

SmartAgent: Extending the JADE Agent Behavior Model

and

Using UML Statemachine Models for More Precise and Flexible JADE Agent
Behaviors

These are available on our Agent Research Group web page:
http://control.cs.utah.edu/PAResearch.html

We are working to try to make this freely available, but it isn't there
yet. Our latest version allowed drawing state machine diagrams in Visio
and then automatically generating the code. The drawing and generation
stuff still could use some improvements, but the basic state machine was
good and solid.

Anyway, the papers (mostly the second one, but understanding the first
one helps with the second one), should give you some more info.

Bob.

-----Original Message-----
From: jade-develop-admin@sharon.cselt.it
[mailto:jade-develop-admin@sharon.cselt.it] On Behalf Of Schneider
Martin
Sent: Monday, October 07, 2002 5:37 AM
To: 'jade-develop'
Cc: Kleegrewe Christian; Olpp Dieter; Mueller Joerg
Subject: [jade-develop] Automatic generation of Finite State Machines

Hi!

We have already worked with the FSMBehaviour.
For each application we have written specific FSMBehaviours
which could not be used otherwise.

Now we want to do something different, more generic:
1. Describe the Finite State Machine in a suitable description language
   (perhaps XML or an XML-derivate)
2. During runtime, the agent should read this description, build
automatically
   the corresponding FSMBehaviour; and then execute it.

We assume that atomic operations are available as JADE behaviours.

Has anybody worked already in this direction and can give us some hints?
Or knows where related work is done?

Thanks!

Martin

PS:
A further assumption could be:
there are mainly three kinds of states:
- receiving (waiting for something)
- if everything is received or a timeout reached, do something with the
received information
- then send something to other agents
- receive again (which can be a different state from the receive-state
above;
   e.g. dependent from the information which the agent waits for.)

=================================================
Martin Schneider
Siemens AG
CT IC 6 Intelligent Autonomous Systems
tel.: +49 89 636-44257 fax.: +49 89 636-41423
e-mail: Martin.Schneider@siemens.com
 

_______________________________________________
jade-develop mailing list
jade-develop@sharon.cselt.it
http://sharon.cselt.it/mailman/listinfo/jade-develop
UNSUBSCRIBE INSTRUCTIONS AT http://jade.cselt.it/mailing.htm#unsubscribe



This archive was generated by hypermail 2a22 : Thu Oct 10 2002 - 17:43:32 MET DST