Re: MathSource Notebook Smith Normal Form Error - How to Correct?
- To: mathgroup at smc.vnet.net
- Subject: [mg131557] Re: MathSource Notebook Smith Normal Form Error - How to Correct?
- From: daniel.lichtblau0 at gmail.com
- Date: Thu, 29 Aug 2013 02:23:03 -0400 (EDT)
- Delivered-to: l-mathgroup@mail-archive0.wolfram.com
- Delivered-to: l-mathgroup@wolfram.com
- Delivered-to: mathgroup-outx@smc.vnet.net
- Delivered-to: mathgroup-newsendx@smc.vnet.net
- References: <kvhino$n1b$1@smc.vnet.net>
On Tuesday, August 27, 2013 1:59:04 AM UTC-5, amzoti wrote: > Hi > > I am trying to do Rational Canonical Forms for matrices and there are no built-in things for this and Smith Normal Forms. > > I want to calculate the rational canonical matrix and the P such that A = PRP^(-1). This is a generalized variant of the Jordan Form (which I know how to do). > > I found this notebook on Math Source from V6: http://library.wolfram.com/infocenter/MathSource/7081/ > > It would be great if this package could show how the invariant canonical decomposition of [xI - A] was done (row and column operations). > > When I download this and try using it, it does not work and I cannot figure out why. > > Is there a way to correct this package so it works in Mathematica Version 9? > > Did I miss some other command in Mathematica V9 that can do all of this stuff? > > Thanks What did you actually try? I show a simple example below that appears to give a plausible result (but it's late and I'm not going to check it). In[81]:= PolynomialSmithForm[x*IdentityMatrix[3] - {{1,2,3},{3,-1,6},{2,7,1}},x] Out[81]= {{{1, 0, 0}, {0, 1, 0}, {0, 0, -44 - 55 x - x^2 + x^3}}, {{{0, -(1/3), 0}, {4/167, 14/ 167, -(23/167) + (2 x)/167}, {23 + 2 x, -3 + 7 x, -7 + x^2}}, {{1, 1/3 + x/3, -(929/501) + (98 x)/501 + (23 x^2)/501 - (2 x^3)/ 501}, {0, 1, 73/167 + (25 x)/167 - (2 x^2)/167}, {0, 0, 1}}}} Mathematica 9 does not have this type of function. I expect there to be a well hidden way to do it in version 10. That is to say, it will not be in System` context. ControlTheory`, perhaps. That's for the polynomial version. I need to check whether we will also have the integer version there. Daniel Lichtblau Wolfram Research