Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!rpi!uupsi!pinet!pat2
From: pat2@pinet.aip.org (patrick_tibbits)
Newsgroups: comp.ai.neural-nets
Subject: Possibility of resource allocation by NN
Keywords: constrained optimization, allocation of resources
Message-ID: <1991Feb1.055546.15511@pinet.aip.org>
Date: 1 Feb 91 05:55:46 GMT
Sender: Patrick Tibbits pat2@pinet.aip.org
Organization: American Institute of Physics
Lines: 8

I would like to train a NN to do allocation of limited assets to maximize
a merit function. Only rub is, the merit function is evaluated thru a 
monte carlo code, making conventional constrained optimization too compute
intensive. Can I feed back the outcome for the MC simulation to train a 
neural net so I won't have to solve the optimization problem for every
new set of input conditions? 

Patrick Tibbits pat2@pinet.aip.org