Date:         Wed, 11 Jul 2001 08:53:18 -0400
Reply-To:     mcallasterd@LEE.ARMY.MIL
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         "McAllaster, Douglas L. LTC" <mcallasterd@LEE.ARMY.MIL>
Subject:      Dijkstra's algorithm?
Comments: To: SAS-L@UGA.CC.UGA.EDU

Sterling,

I'm surprised.
I've solved pure netflow problems with 150,000 nodes very quickly.
It's always worked well for me.
Could you have inadvertently included a non-arc variable?
Side constraints slow the network simplex considerably.

Doug
LTC Doug McAllaster
McAllasterD@Lee.Army.Mil


Date:    Wed, 11 Jul 2001 12:06:44 GMT
From:    Sterling Price <ssprice@COX-INTERNET.COM>
Subject: Re: Dijkstra's algorithm?

Thanks for the reply, Doug.  Unfortunately I've tried the problem using Proc
Netflow and it didn't seem to scale very well -- it worked OK for a small
sample problem,  but can't seem to handle the "real world" problem I need it
for.  Maybe I'm fooling myself to think coding it by hand in SAS would be
any
better,  but the alternative seems to be C or C++, which I'd rather not get
into if it can be done in SAS.

Regards,

Sterling Price