wu :: forums
« wu :: forums - Constant Norm »

Welcome, Guest. Please Login or Register.
Nov 28th, 2024, 12:42pm

RIDDLES SITE WRITE MATH! Home Home Help Help Search Search Members Members Login Login Register Register
   wu :: forums
   riddles
   putnam exam (pure math)
(Moderators: Grimbal, SMQ, Icarus, towr, william wu, Eigenray)
   Constant Norm
« Previous topic | Next topic »
Pages: 1  Reply Reply Notify of replies Notify of replies Send Topic Send Topic Print Print
   Author  Topic: Constant Norm  (Read 482 times)
william wu
wu::riddles Administrator
*****





   
WWW

Gender: male
Posts: 1291
Constant Norm  
« on: Feb 8th, 2006, 8:44pm »
Quote Quote Modify Modify

Consider the matrix differential equation

x' = A x

where x is a state vector in Rn, x' is the derivative of that vector with respect to time, and A is an n-by-n matrix. Find conditions on A which guarantee that the Euclidean norm || x ||2 remains constant.
 
Source: Stephen Boyd, Stanford EE Quals 2006
« Last Edit: Feb 9th, 2006, 1:13am by william wu » IP Logged


[ wu ] : http://wuriddles.com / http://forums.wuriddles.com
Eigenray
wu::riddles Moderator
Uberpuzzler
*****






   


Gender: male
Posts: 1948
Re: Constant Norm  
« Reply #1 on: Feb 8th, 2006, 9:50pm »
Quote Quote Modify Modify

o(n)?  (That is, the Lie algebra of O(n), i.e., skew-symmetric matrices.)
« Last Edit: Feb 8th, 2006, 9:54pm by Eigenray » IP Logged
towr
wu::riddles Moderator
Uberpuzzler
*****



Some people are average, some are just mean.

   


Gender: male
Posts: 13730
Re: Constant Norm  
« Reply #2 on: Feb 9th, 2006, 12:50am »
Quote Quote Modify Modify

Which norm is that? ||x||2 (i.e euclidean distance?)
IP Logged

Wikipedia, Google, Mathworld, Integer sequence DB
william wu
wu::riddles Administrator
*****





   
WWW

Gender: male
Posts: 1291
Re: Constant Norm  
« Reply #3 on: Feb 9th, 2006, 1:21am »
Quote Quote Modify Modify

Yes sorry, Euclidean norm.  
 
That's right Eigenray. I thought it was a cute result Smiley
IP Logged


[ wu ] : http://wuriddles.com / http://forums.wuriddles.com
Eigenray
wu::riddles Moderator
Uberpuzzler
*****






   


Gender: male
Posts: 1948
Re: Constant Norm  
« Reply #4 on: Feb 9th, 2006, 9:14am »
Quote Quote Modify Modify

To elaborate, we have
x(s) = esAx,
so we want esA to be orthonormal for all s, which is precisely the statement that A is in the Lie algebra o(n) associated to the Lie group O(n).
 
esA is orthonormal when
I = (esA)t esA = esA^t esA.
Differentiating wrt s, and since At commutes with eA^t, we have
esA^t (At+A) esA = 0,
and so we find At+A = 0.  For sufficiency, run the argument backwards, or just note that eA is clearly orthonormal when At = -A: since A and -A commute,
(esA)t esA = e-sAesA = eO = I.
 
By the same argument, if we had x in Cn, we'd want A to be in u(n), the Lie algebra of U(n), i.e., skew-Hermitian.
IP Logged
william wu
wu::riddles Administrator
*****





   
WWW

Gender: male
Posts: 1291
Re: Constant Norm  
« Reply #5 on: Feb 10th, 2006, 2:48am »
Quote Quote Modify Modify

Another way:
 
- Want || x || to be constant, so, the derivative of ||x|| with respect to time should be zero. That is,
 
0 = (d/dt) [ || x || ] = (d/dt) [ sqrt( xT x ) ]
 
 
- Just throw away the square root. Thus we need  
 
0 = (d/dt) [ xT x ]
= 2 xT x'
= 2 xT A x
 
So 0 = xT A x.  
 
 
- Recall  xT A x = xT ((A + AT)/2) x. So when A = -AT, the quadratic form is zero.
IP Logged


[ wu ] : http://wuriddles.com / http://forums.wuriddles.com
Pages: 1  Reply Reply Notify of replies Notify of replies Send Topic Send Topic Print Print

« Previous topic | Next topic »

Powered by YaBB 1 Gold - SP 1.4!
Forum software copyright © 2000-2004 Yet another Bulletin Board