Path: utzoo!censor!geac!torsqnt!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!usc!apple!agate!darkstar!helmut.scs.carleton.ca
From: mbm@helmut.scs.carleton.ca (Max B. Maklin)
Newsgroups: comp.os.research
Subject: Request for learning techniques in distributed systems
Message-ID: <10253@darkstar.ucsc.edu>
Date: 18 Dec 90 21:16:37 GMT
Sender: usenet@darkstar.ucsc.edu
Organization: Carleton University
Lines: 30
Approved: comp-os-research@jupiter.ucsc.edu


I am posting in hope of finding relevant sources to the problem of
learning intelligent load balancing in a distributed system.  The
system may be either homogeneous or hetergeneous, but I am inclined
to hetergeneous systems.  I am interested in any readings, (i.e.,
tech reports, journals, articles, thesis, dissertions, etc.) on the
application of sub-symbolic learning to adaptive, dynamic 
load-balancing in distributed systems, especially those which are 
loosely-coupled systems (i.e., communication costs are of importance).
The types of learning techniques I would like to employ include
neural networks, learning automata, and genetic algorithms, solely, 
or in conjuction, but would be interested to hear from others who have
used different techniques.

I would also be interested in hearing from any other fellow researchers
interested in similar notions for possible correspondance and exchange 
of results.

I would appreciate that all replys be sent to my address and if there
is enough response, I would be happy to post it to the relevent news
groups at a later time.


-- 
+-----------------------------------------------------------------+
|  Max Maklin                 |                                   |
|  School of Computer Science |  e-mail: mbm@hans.scs.carleton.ca |
|  Carleton University        |                                   |
|  Ottawa, CAN K1S 5B6        |                                   |
+-----------------------------------------------------------------+