Date: Fri, 16 Dec 2005 16:56:09 -0800
Reply-To: David L Cassell <davidlcassell@MSN.COM>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: David L Cassell <davidlcassell@MSN.COM>
Subject: Re: Help with PROC LP, integer program initial solution
Content-Type: text/plain; format=flowed
>Thanks for the tips. The problem is IP, with some integer and some
>binary variables. I'm not familiar with the term "driving
>coefficients", I've been using the BINARY and INTEGER types of row to
>specify the variables.
>The _ORLP_ macro variable isn't telling me anything interesting.
>I've tried using the PRIMALIN method, but I'm having trouble putting my
>initial solution into the right format, if anyone has any tricks on
>that let me know. It's years since I've done any math programming, so
>trying to find which variables are basic and which are just at their
>upper bound is beyond me. I tried creating a new problem which set all
>the variables to their values in the optimal solution, and using the
>PRIMALOUT from that as a PRIMALIN to a modified problem where the new
>constraint is not binding, but it just told me that the basis in the
>PRIMALIN data set is singular! Not quite sure why that is or what it
>I'd really prefer to use the BASIC type if I could get that working...
>it seems to be the option designed for this purpose.
Hmmm. You took the PRIMALOUT= data set, and fed it in as a new PRIMALIN=
data set, and PROC LP told you that the basis was singular? Did the &_ORLP_
macro variable tell you the same thing?
Is it possible you have over-specified your model?
David L. Cassell
3115 NW Norwood Pl.
Corvallis OR 97330
Express yourself instantly with MSN Messenger! Download today - it's FREE!