5.2-exam2-2012-strang¶
Quiz 2, 2012 (strang) · Section 5.2: Computing and Using Determinants Source: exam2_2.pdf, pp. 31-34
Problem¶
2 (30 pts.)
(a) Suppose the matrix \(A\) factors into \(A = PLU\) with a permutation matrix \(P\), and 1's on the diagonal of \(L\) (lower triangular) and pivots \(d_1, \dots, d_n\) on the diagonal of \(U\) (upper triangular). What is the determinant of \(A\)? EXPLAIN WHAT RULES YOU ARE USING.
(b) Suppose the first row of a new matrix \(A\) consists of the numbers 1, 2, 3, 4. Suppose the cofactors \(C_{ij}\) of that first row are the numbers 2, 2, 2, 2. (Cofactors already include the \(\pm\) signs.) Which entries of \(A^{-1}\) does this tell you and what are those entries?
(c) What is the determinant of the matrix \(M(x)\)? For which values of \(x\) is the determinant equal to zero?
Solution
(a) Use \[ \det(A) = \det(P) \cdot \det(L) \cdot \det(U), \] where we make two uses of the rule \(\det(MN) = \det(M)\det(N)\), for any two \(n \times n\) matrices \(M\) and \(N\). The determinant of a triangular matrix is the product of its diagonal entries, so \[ \det(L) = 1 \quad \text{and} \quad \det(U) = d_1d_2 \cdots d_n. \] The determinant changes sign whenever two rows are swapped, so \[ \det(P) = \begin{cases} +1 & \text{if } P \text{ is even (even number of row exchanges)} \\ -1 & \text{if } P \text{ is odd (odd number of row exchanges)} \end{cases} \] and so \(\det(A) = \pm d_1d_2 \cdots d_n\), where the sign depends on the parity of \(P\). (b) Using the cofactor expansion in the first row gives \[ \det(A) = a_{11}C_{11} + a_{12}C_{12} + a_{13}C_{13} + a_{14}C_{14} = 1 \cdot 2 + 2 \cdot 2 + 3 \cdot 2 + 4 \cdot 2 = 20. \] As \(A^{-1} = C^T/\det(A)\), where \(C\) is the cofactor matrix, this data gives us the entries of the first column of \(A^{-1}\); they are all \(2/20 = 1/10\). (c) Solution no. 1. From, for instance, the big formula, we know that \(\det(M)\) is a cubic polynomial in \(x\). Say \(\det(M) = ax^3 + bx^2 + cx + d\). We can calculate \(d\) by setting \(x = 0\). Using the cofactor expansion in the last column, we get \[ d = -\begin{vmatrix} 1 & -1 & 2 \\ 1 & 1 & 4 \\ 1 & -1 & 8 \end{vmatrix} = -\begin{vmatrix} 1 & -1 & 2 \\ 0 & 2 & 2 \\ 0 & 0 & 6 \end{vmatrix} = -12. \] We will determine the other coefficients by finding three roots of \(\det(M)\). \(x\) is a root of \(\det(M)\) if and only if \(M(x)\) is singular. Now notice that \[ (1, 1, 1) = (x, x^2, x^3) \ \text{for } x = 1, \qquad (1, -1, 1) = (x, x^2, x^3) \ \text{for } x = -1, \qquad (2, 4, 8) = (x, x^2, x^3) \ \text{for } x = 2. \] Thus \(M(x)\) is singular for \(x = 1, -1\), and 2; moreover, this implies that \[ \det(M) = a(x - 1)(x + 1)(x - 2). \] As \(d = 2a\), we must have \(a = -6\). Thus \[ \det(M) = -6(x - 1)(x + 1)(x - 2) = -6x^3 + 12x^2 + 6x - 12. \] The values of \(x\) for which \(M(x)\) is singular are 1, \(-1\), and 2. Solution no. 2. \[ \begin{vmatrix} 1 & 1 & 1 & 1 \\ 1 & -1 & 2 & x \\ 1 & 1 & 4 & x^2 \\ 1 & -1 & 8 & x^3 \end{vmatrix} = \begin{vmatrix} 1 & 1 & 1 & 1 \\ 0 & -2 & 1 & x - 1 \\ 0 & 0 & 3 & x^2 - 1 \\ 0 & -2 & 7 & x^3 - 1 \end{vmatrix} = \begin{vmatrix} 1 & 1 & 1 & 1 \\ 0 & -2 & 1 & x - 1 \\ 0 & 0 & 3 & x^2 - 1 \\ 0 & 0 & 6 & x^3 - x \end{vmatrix} \] \[ = \begin{vmatrix} -2 & 1 & x - 1 \\ 0 & 3 & x^2 - 1 \\ 0 & 6 & x^3 - x \end{vmatrix} = -2 \begin{vmatrix} 3 & x^2 - 1 \\ 6 & x^3 - x \end{vmatrix} = -6x^3 + 12x^2 + 6x - 12. \] In the first step, subtract the first row from the second, third, and fourth rows. In the second step, subtract the second row from the fourth. For the third and fourth steps, use the cofactor expansion in the first column. We factorize \(\det(M)\) by guessing roots, trying small integers; we find that 1, \(-1\), and 2 are all roots, which gives \[ \det(M) = -6(x - 1)(x + 1)(x - 2). \] The values of \(x\) for which \(M(x)\) is singular are 1, \(-1\), and 2.Keywords: determinant, LU factorization, cofactors, inverse matrix, polynomial