Coupled eigenvalue-like problem

  • Thread starter Thread starter Office_Shredder
  • Start date Start date
  • Tags Tags
    Coupled
Click For Summary
SUMMARY

The discussion centers on proving the existence of an invertible matrix P such that PtAP equals the identity matrix and PtBP equals a diagonal matrix with distinct non-zero entries a, b, and c, derived from the determinant equation det(xA-B)=0. The matrices A and B are complex symmetric 3x3 matrices, and the proof hinges on the existence of three distinct vectors corresponding to the eigenvalues. The initial attempts to construct P using the vectors x, y, and z highlight the challenges in ensuring the distinctness of the eigenvalues while maintaining the properties of the matrices involved.

PREREQUISITES
  • Understanding of complex symmetric matrices
  • Knowledge of eigenvalue problems and their properties
  • Familiarity with determinants and their implications in linear algebra
  • Experience with matrix diagonalization techniques
NEXT STEPS
  • Study the properties of complex symmetric matrices and their eigenvalues
  • Learn about matrix diagonalization and the conditions for invertibility
  • Explore the implications of the determinant in linear transformations
  • Investigate advanced linear algebra techniques for constructing invertible matrices
USEFUL FOR

Mathematicians, graduate students in linear algebra, and anyone involved in theoretical physics or engineering applications requiring a deep understanding of matrix theory and eigenvalue problems.

Office_Shredder
Staff Emeritus
Science Advisor
Gold Member
Messages
5,706
Reaction score
1,592

Homework Statement


If A, B are (complex) symmetric 3x3 matrices and det(xA-B) has three distinct solutions a,b and c, prove there exists P invertible such that

PtAP = I the identity matrix
PtBP = diag(a,b,c) the diagonal matrix with non-zero entries a,b and c

Homework Equations



det(xA-B)=0 iff there exists v =/= 0 s.t. xAv = Bv

The Attempt at a Solution



So we have three distinct vectors we'll call x,y and z such that

aAx = Bx
bAy = By
cAz = Bz

and from here we want to construct P. The first thing I notice is that Ax, Ay and Az are all non-zero, as otherwise Bx (or y or z) is too and one of a,b or c can be arbitrary (and hence they are not necessarily distinct). But then just because a, b and c aren't necessarily distinct doesn't mean that there aren't distinct a, b and c that do satisfy the condition. This way lies madness, so I decided to go another route:

If we naively try P = [x y z] then AP = [Ax Ay Az] and BP = [aAx bAy cAz] which looks like a pretty good start, but then Pt doesn't do anything useful unless I'm missing something big.

Basically I'm just looking for what the right starting off point is
 

Similar threads

Replies
1
Views
1K
  • · Replies 6 ·
Replies
6
Views
2K
  • · Replies 5 ·
Replies
5
Views
6K
  • · Replies 2 ·
Replies
2
Views
4K
Replies
7
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
1
Views
1K
  • · Replies 5 ·
Replies
5
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K