Date:         Mon, 7 Oct 1996 10:01:00 -0400
Reply-To:     Jin Qian <Jin.Qian@MCI.COM>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Jin Qian <Jin.Qian@MCI.COM>
Subject:      Re: regression w/ nonlinear constraints
Comments: To: chris bratcher <bratcher@jeeves.la.utexas.edu>

Chris:
 
Try PROC MODEL.  It allows you to explicitly specify parameter
restrictions in almost any form.
 
Jin Qian
COMSYS Technical Service, Inc.
@ MCI.
 
 
-------- REPLY, Original message follows --------
Date: Sat, 5 Oct 96 07:53 EST
From: chris bratcher <bratcher@jeeves.la.utexas.edu>
To: Multiple recipients of list SAS-L <SAS-L@vm.marist.edu>
Subject: regression w/ nonlinear constraints
Message-Id: <30961005125303/0002127673DF3EM@MCIMAIL.COM>
 
Any help would be appreciated.  I'm trying to estimate a variation of a
partial-adjustment model that contains nonlinear constraints on some of
the parameters.
 
The model is:
 
Yi = A  + B1*X1  + B2*X2 + B3*X3 + . . .
 
For some rather complicated reasons I need to constrain the parameters
as follows:
 
A/(1 - B1) = B2/B3
 
I've done a quick run-through of the SAS manuals and found some material
on linear constraints but nothing on NONLINEAR ones such as these.  Any
suggestions?
 
Thanks,
 
Chris Bratcher
University of Texas
-------- REPLY, End of original message --------