jason-users Mailing List for jason

Hello Jomi...

I get this version in the repository and I try use the .send internal
action with multiples destiny. The .send does not work :/. (Similar with
the previous error).

thanks...


-- 
Ass: Marcos Côrtes
Mestrando Computação - UFF

Hi Jason's develops...

I have a question about the Agent class.

A derived class can access the BeliefBase in all methods because there is
the method getBB(). And the user can access it in any method of a derived
Agent class.

Well, can the user access the Event Base and the Intention Base in any
method of a derived class by the method getTS().getC()? Is it a "good
pattern"? (is it common or will I bring me a lot of problems?)


I ask it because I has implement a Emotional Agent Framework based in Jason
and I need define that bases I has access in one extended method of class
Agent.

Thanks...

-- 
Ass: Marcos Côrtes
Mestrando Computação - UFF

Dear Marcos,

you may try to increase the timeout for agent's action (see for an example the code of FoodEnvironment.java).

Jomi

On 19/01/2012, at 18:11, Marcos Côrtes wrote:

> Hi jason's develops
> 
> 
> I has tried use the FoodSimulation example to test my emotional agent.
> 
> well, the agents in this example are reactive because their plans execute an action there elicited.
> 
> However, to my system I need deliberative agents, with goals and plans concurrence, because I use to the emotional mechanism.
> 
> but, I has tried implement a .asl with it. but the experiment has tilt, because the agent has not executed a action in the end of cicle.
> 
> I looked for the problem in the code, and I found the SteeppedEnvironment gives to the agent only a "time frame" to it execute a action, after the deadline, the environment will ignore the action.
> 
> Because it, a slow agent cannot run a action because the environment does not wait it.
> 
> I changed the code to the environment does not ignore the action, but I do not have success.
> 
> 
> Do You have any suggestion for me? I would use a normal environment, but I think it will be harder and the system will be different of original.
> 
> Thanks...
> -- 
> Ass: Marcos Côrtes
> Mestrando Computação - UFF
> ------------------------------------------------------------------------------
> Keep Your Developer Skills Current with LearnDevNow!
> The most comprehensive online learning library for Microsoft developers
> is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3,
> Metro Style Apps, more. Free future releases when you subscribe now!
> http://p.sf.net/sfu/learndevnow-d2d_______________________________________________
> Jason-users mailing list
> Jas...@li...
> https://lists.sourceforge.net/lists/listinfo/jason-users

-- 
Jomi Fred Hubner
Federal University of Santa Catarina
Department of Automation and Systems Engineering
PO Box 476, Florianópolis, SC
88040-900 Brazil
http://www.das.ufsc.br/~jomi

I used the Oracle profile that is in the JDK called Visual VM.

http://docs.oracle.com/javase/6/docs/technotes/guides/visualvm/profiler.html



It did help me with my application. I has a lot of call to sleep() in my
agents because the communication protocol (I used a simple contract-net).

-- 
Ass: Marcos Côrtes
Mestrando Computação - UFF

Hi jason's develops


I has tried use the FoodSimulation example to test my emotional agent.

well, the agents in this example are reactive because their plans execute
an action there elicited.

However, to my system I need deliberative agents, with goals and plans
concurrence, because I use to the emotional mechanism.

but, I has tried implement a .asl with it. but the experiment has tilt,
because the agent has not executed a action in the end of cicle.

I looked for the problem in the code, and I found the SteeppedEnvironment
gives to the agent only a "time frame" to it execute a action, after the
deadline, the environment will ignore the action.

Because it, a slow agent cannot run a action because the environment does
not wait it.

I changed the code to the environment does not ignore the action, but I do
not have success.


Do You have any suggestion for me? I would use a normal environment, but I
think it will be harder and the system will be different of original.

Thanks...
-- 
Ass: Marcos Côrtes
Mestrando Computação - UFF

Dear Vin,

As far as I know, there isn't so much on profiling tools specific for agents. (I guess it is an open issue yet). Regarding the performance of AOP languages, Francisco Grimaldo has published some papers on the subject (using Jason). I guess also Koen Hindriks and Tristan Behrens have done some work on that also.

It follows some alternatives I have used to workaround with this problem.

1. use normal java profiling tools. In eclipse, for instance, run the class RunCentralisedMAS (with your project as parameter) under profiling. 

2. use custom classes (agent, arch, belief base, ...) to count number of messages, perception, etc. You find attached an example.

HTH,

Jomi