6.3-final-2008-strang-b¶
Final Exam, 2008 (strang) · Section 6.3: Symmetric Positive Definite Matrices Source: finalexam_1.pdf, p. 220
Problem¶
9 (7 pts.) (a) Find the inverse \(L^{-1}\) of this real triangular matrix \(L\):
\[
L = \begin{bmatrix} 1 & 0 & 0 \\ a & 1 & 0 \\ 0 & a & 1 \end{bmatrix}
\]
You can use formulas or Gauss-Jordan elimination or any other method.
(b) Suppose \(D\) is the real diagonal matrix \(D = \text{diag}(d, d^2, d^3)\). What are the conditions on \(a\) and \(d\) so that the matrix \(A = LDL^T\) is (three separate questions, one point each)
(i) invertible? (ii) symmetric? (iii) positive definite?
Solution
(a) I'll do Gauss-Jordan elimination. \[ \begin{bmatrix} 1 & 0 & 0 & 1 & 0 & 0 \\ a & 1 & 0 & 0 & 1 & 0 \\ 0 & a & 1 & 0 & 0 & 1 \end{bmatrix} \rightsquigarrow \begin{bmatrix} 1 & 0 & 0 & 1 & 0 & 0 \\ 0 & 1 & 0 & -a & 1 & 0 \\ 0 & a & 1 & 0 & 0 & 1 \end{bmatrix} \] \[ \rightsquigarrow \begin{bmatrix} 1 & 0 & 0 & 1 & 0 & 0 \\ 0 & 1 & 0 & -a & 1 & 0 \\ 0 & 0 & 1 & a^2 & -a & 1 \end{bmatrix} \] (b) Note that \(L\) is invertible no matter what \(a\) is, and \(D\) is invertible so long as \(d \neq 0\). So \(A = LDL^T\) will be invertible whenever \(d \neq 0\). If \(d = 0\), then of course \(A\) can't be invertible. The matrix \(A\) is always symmetric, since \[ A^T = (LDL^T)^T = LD^TL^T = LDL^T. \] Because \(A\) is always symmetric, to check positive-definiteness we just need to check that the pivots are all positive. But \(A = LDL^T\) is the "pivot" decomposition for \(A\). So the pivots of \(A\) are \(d, d^2, d^3\), and we need \(d > 0\).Keywords: matrix inverse, triangular matrix, positive definite, LDL factorization