LISTSERV mailing list manager LISTSERV 16.0

Help for ECJ-INTEREST-L Archives


ECJ-INTEREST-L Archives

ECJ-INTEREST-L Archives


ECJ-INTEREST-L@LISTSERV.GMU.EDU


View:

Message:

[

First

|

Previous

|

Next

|

Last

]

By Topic:

[

First

|

Previous

|

Next

|

Last

]

By Author:

[

First

|

Previous

|

Next

|

Last

]

Font:

Proportional Font

LISTSERV Archives

LISTSERV Archives

ECJ-INTEREST-L Home

ECJ-INTEREST-L Home

ECJ-INTEREST-L  October 2006

ECJ-INTEREST-L October 2006

Subject:

Any interest in another type of master-slave?

From:

Robert Baruch <[log in to unmask]>

Reply-To:

ECJ Evolutionary Computation Toolkit <[log in to unmask]>

Date:

Thu, 26 Oct 2006 13:50:36 -0400

Content-Type:

text/plain

Parts/Attachments:

Parts/Attachments

text/plain (86 lines)

Hi all,

I've had an idea for a few months now, that I've been letting  
percolate in the back half of my brain. If I understand the way the  
current master-slave processing works, the master starts up, and  
uniformly queues jobs to slaves. Each job consists of one or more  
individuals (depending on whether we're running a SimpleProblemForm  
or a GroupedProblemForm).

For each job on the queue, the master sends the appropriate  
individuals to the slave, the slave evaluates them, and then sends  
the results to the master.

So for a SimpleProblemForm with, say, 1024 individuals and 4 slaves,  
there are 256 cycles of send-wait-result per slave.

My concerns are:

1. The algorithm doesn't address the relative processing power of  
each slave in any way.
2. As the comments in the API state, master-slave only makes sense if  
the communication time for a single job is a fraction of the  
evaluation time.


To address these issues, I propose to work on a solution that will at  
least work for SimpleProblemForms, as follows:

INIT: For each slave, enter 1.0 in efficiency array.

1. Master divides individuals to be evaluated among slaves according  
to their efficiency as follows:

#individuals_to_evaluate[slave] = e[slave] / sum(e)

where e[slave] is the efficiency entry in the array, and
sum(e) is the sum of all entries in the efficiency array.

2. Send ALL individuals to be evaluated to each slave. Master awaits  
results.
3. Slave evaluates all individuals sent to it.
4. Slave sends ALL results to master.

5. Once all results are in, based on the times and number of  
individuals evaluated, master updates the slave efficiency array as  
follows:

e[slave] = #individuals_evaluated[slave] / time_to_evaluate[slave]


6. Master computes next generation. Repeat to step 1.



I feel this addresses the two issues in this way:

1. Although not perfect, each slave, after the first generation, is  
assigned a number of individuals  to evaluate such that all slaves  
should return at statistically the same time, give or take.

2. The communication time becomes small relative to the evaluation  
time by having the slaves evaluate many individuals per communication  
rather than just one individual.



One issue is what happens if a slave fails. In this case, it would  
probably make sense to have a timeout per slave, such that if the  
slave does not return within, say, twice the expected time, the slave  
is considered nonresponsive, and removed from the list of available  
slaves.

The expected time would be calculated based on the time it takes for  
the first slave to return. In generations > 1 this makes sense  
because all slaves are expected to return in roughly the same amount  
of time. For generation 1, this will kill off slaves who are  
unreasonably slow relative to the fastest slave. This isn't too good  
if the fastest slave happens to be an outlier, though.


Any thoughts? Comments? Criticisms?

Thanks,

--Rob

Top of Message | Previous Page | Permalink

Advanced Options


Options

Log In

Log In

Get Password

Get Password


Search Archives

Search Archives


Subscribe or Unsubscribe

Subscribe or Unsubscribe


Archives

September 2019
August 2019
June 2019
April 2019
March 2019
January 2019
December 2018
November 2018
October 2018
July 2018
May 2018
January 2018
December 2017
November 2017
October 2017
September 2017
August 2017
July 2017
June 2017
May 2017
April 2017
March 2017
February 2017
January 2017
December 2016
November 2016
October 2016
September 2016
August 2016
July 2016
June 2016
May 2016
April 2016
March 2016
February 2016
January 2016
December 2015
November 2015
October 2015
September 2015
August 2015
July 2015
June 2015
May 2015
April 2015
March 2015
February 2015
January 2015
December 2014
November 2014
September 2014
August 2014
July 2014
June 2014
May 2014
April 2014
March 2014
February 2014
January 2014
December 2013
November 2013
October 2013
September 2013
August 2013
July 2013
June 2013
May 2013
April 2013
March 2013
February 2013
January 2013
December 2012
November 2012
October 2012
September 2012
July 2012
June 2012
May 2012
April 2012
March 2012
February 2012
January 2012
December 2011
November 2011
October 2011
September 2011
August 2011
July 2011
June 2011
May 2011
April 2011
March 2011
February 2011
January 2011
December 2010
November 2010
October 2010
September 2010
August 2010
July 2010
June 2010
May 2010
April 2010
March 2010
February 2010
January 2010
December 2009
November 2009
October 2009
September 2009
August 2009
July 2009
June 2009
May 2009
April 2009
March 2009
February 2009
January 2009
December 2008
November 2008
October 2008
September 2008
August 2008
July 2008
June 2008
May 2008
April 2008
March 2008
February 2008
January 2008
December 2007
November 2007
October 2007
September 2007
August 2007
July 2007
June 2007
May 2007
April 2007
March 2007
February 2007
January 2007
December 2006
November 2006
October 2006
September 2006
August 2006
July 2006
June 2006
May 2006
April 2006
March 2006
February 2006
January 2006
December 2005
November 2005
October 2005
September 2005
August 2005
July 2005
June 2005
May 2005
April 2005
March 2005
February 2005
January 2005
December 2004
November 2004
September 2004
August 2004
July 2004
June 2004
May 2004
April 2004
March 2004

ATOM RSS1 RSS2



LISTSERV.GMU.EDU

CataList Email List Search Powered by the LISTSERV Email List Manager