MASON-INTEREST-L Archives

May 2014

MASON-INTEREST-L@LISTSERV.GMU.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Russell Thomas <[log in to unmask]>
Reply To:
MASON Multiagent Simulation Toolkit <[log in to unmask]>
Date:
Wed, 14 May 2014 21:00:47 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (46 lines)
Generally, you need to remove the "destroyed" agents from ALL data
structures where thy might be referenced-- grids, bags, ArrayLists, and so
on.  Only then will the Java garbage collector remove them from memory.

Russ


On 5/14/14 8:43 PM, "Sean Luke" <[log in to unmask]> wrote:

> SparseGrid2D contains objects.  It is a representation of space as an infinite
> grid.
> 
> The Schedule contains "agents", which MASON calls Steppables.  It is a
> representation of time.
> 
> It sounds like you have added your agent object to both data structures.
> 
> If you have stopped an agent in the Schedule (after, I presume, scheduling it
> repeating), it will eventually drop out of the Schedule.  Why would you think
> it'd also be deleted from the SparseGrid2D?  Or any other data structure?
> 
> When you call stop() on your stopper, you need to also manually call
> removeObject() on your SparseGrid2D.
> 
> Sean
> 
> On May 14, 2014, at 11:30 PM, Francesco Pizzitutti
> <[log in to unmask]> wrote:
> 
>> Hi,
>> 
>> I am creating and destroying a lot of agents during my simulation. They are
>> all 
>> attached to a SparseGrid2D. For some reason when I stop, with the stopper an
>> agent, the reference between the agent and the grid remain active. This leads
>> to 
>> an increase in the memory use and to a progressive slowdown of the
>> simulation.  
>> 
>> If I periodically reset the sparseGrid2D containing all the agents this
>> remove all 
>> the references and free memory space. Is this the only solution to this
>> problem?
>> 
>> Francesco

ATOM RSS1 RSS2