Finding SPD Matrix: Vanderberghe and Boyd Reader 6.4 Solution

  • Thread starter earlh
  • Start date
  • Tags
    Homework
In summary, the conversation discusses a problem from a text on linear algebra where the goal is to show when a matrix is SPD (symmetric positive definite). The problem involves finding an interval for a scalar, a, so that a specific matrix, B, is SPD. The conversation also includes an attempt at a solution using a vector, v, and its transpose. The conversation ends with a question about the meaning of SPD.
  • #1
earlh
12
0
Hi -- so, I'm working through a text on my own 5 years out of college having completely forgotten more or less all linear algebra. Here's my problem -- 6.4 from a reader by Vanderberghe and Boyd which is btw excellent.

Homework Equations


I have to show when a matrix is SPD. Setup:
let [tex]A[/tex] which is n by n be SPD. For what values of scalar [tex]a[/tex] is the matrix
[tex] B = \left[ \begin{array}{cc}
A & ae_{1} \\
ae_{1}^{T} & 1 \\
\end{array} \right] [/tex]
where [tex] e_{1} [/tex] is the first vector in the usual basis

SPD? I need to find an interval for [tex]a[/tex] so that [tex]B[/tex] is SPD

The Attempt at a Solution


So I let [tex] v = \left[ \begin{array}{c} x \\ y \end{array} \right] [/tex]
where [tex] x [/tex] is (n, 1) and [tex]y [/tex] is 1 by 1.

Multiplying things out, you get
[tex]v^{T} B v = [ x^{t} y ] \left[ \begin{array}{cc}
A & e_{1} \\
e_{1}^{T} & a \\
\end{array} \right] \left[ \begin{array}{c} x \\ y \end{array} \right] [/tex]

[tex] ... = x^{T} A x + y^{2} + 2ay x^{T}e_{1} = x^{T}Ax + y^{2} + 2ayx_{1}[/tex] where [tex] x_{1} [/tex] is the first term in [tex]x[/tex]
Obviously the first 2 terms are greater than or equal to zero, but I'm having trouble figuring out what [tex]a[/tex] I use so that [tex]v[/tex] can be arbitrary...Thanks in advance for any help
 
Last edited:
Physics news on Phys.org
  • #2
Just to satisfy my curiosity, what does SPD stand for?
 
  • #3
vela said:
Just to satisfy my curiosity, what does SPD stand for?

Symmetric positive definite -- sorry for using jargon
 

FAQ: Finding SPD Matrix: Vanderberghe and Boyd Reader 6.4 Solution

What is the Vanderberghe and Boyd Reader 6.4 Solution for finding SPD matrix?

The Vanderberghe and Boyd Reader 6.4 Solution is a method for finding a symmetric positive definite (SPD) matrix, which is a square matrix with all positive eigenvalues. This solution involves solving a convex optimization problem to find the nearest SPD matrix to a given input matrix.

Why is it important to find SPD matrices?

SPD matrices have many applications in fields such as machine learning, signal processing, and physics. They are also used in optimization problems as they have well-behaved properties that make them easier to work with. Therefore, finding SPD matrices accurately and efficiently is crucial in many scientific and engineering applications.

How does the Vanderberghe and Boyd Reader 6.4 Solution work?

The Vanderberghe and Boyd Reader 6.4 Solution involves using a projection onto the set of SPD matrices to find the nearest SPD matrix to a given input matrix. This projection is achieved by solving a convex optimization problem using an iterative algorithm, which converges to the nearest SPD matrix. The solution also includes a stopping criterion to ensure the accuracy of the result.

Can the Vanderberghe and Boyd Reader 6.4 Solution be applied to any matrix?

No, the Vanderberghe and Boyd Reader 6.4 Solution can only be applied to symmetric matrices. This is because the set of SPD matrices is a subset of the set of symmetric matrices. Therefore, the input matrix must be symmetrical for the solution to work.

Are there any limitations to the Vanderberghe and Boyd Reader 6.4 Solution?

One limitation of this solution is that it may not work for matrices with complex eigenvalues. In such cases, other methods must be used. Additionally, the solution may not always converge, which could be due to the input matrix not being close enough to an SPD matrix or the chosen stopping criterion being too strict. It is important to carefully choose the input matrix and the stopping criterion to ensure the accuracy of the result.

Back
Top