Xref: utzoo sci.chem:2344 comp.theory:1178
Path: utzoo!attcan!uunet!cs.utexas.edu!wuarchive!tripos!rick
From: rick@tripos.com (Rick Moll)
Newsgroups: sci.chem,comp.theory
Subject: mapping surfaces into discrete representations
Message-ID: <1990Oct30.202723.8244@tripos.com>
Date: 30 Oct 90 20:27:23 GMT
Distribution: sci,comp
Organization: Tripos Associates, Inc.
Lines: 20


I'm currently working on the design of a system that maps multi-dimensional 
surfaces into a discrete representation.  In particular, it is mapping 
the energy surface of a polymer's repeat-unit into a rotational state 
representation (RIS).  For those without a chemistry background I'm 
essentially attempting to represent a surface as a discrete representation 
by breaking the surface up into states.  By finding "features" of the 
surface, in this case the minima of the surface (more or less), I then map 
these "features" into states.  I start by making a trivial state represen-
tation and then collapse these trivial states into a larger ones.  The 
final result of this mapping is then feed into a RIS property calculation 
system.  The state mapping is thus used to predict the physical properties 
of polymers, using a statistical tool called RIS.  I have had a fair amount 
of success with my ad-hoc pattern matching approach.  But, I am interested 
in knowing if there is some well know algorithms that may be of use to me.  

Thanks,

     Rick Moll
     uunet!tripos!rick