2.3-exam1-2014¶
Exam 1, 2014 · Section 2.3: Matrix Computations and A = LU Source: exam1_2.pdf, pp. 59, 65
Problem¶
Problem 1. Let \(A = \begin{bmatrix} 1 & 1 & 1 \\ 2 & 4 & 4 \\ 3 & 7 & 10 \end{bmatrix}\). (a) Find the \(A = LU\) factorization of the matrix \(A\). (b) Solve the system \(Ax = (3, 10, 20)^T\). Let \(B = \begin{bmatrix} 1 & 1 & 1 \\ 2 & 4 & 4 \\ 3 & 7 & k \end{bmatrix}\) (obtained by replacing the bottom right entry of \(A\) by parameter \(k\)). (c) For which values of \(k\) is the matrix \(B\) singular? (d) Find all values of \(k\) for which the system \(Bx = (1, 2, 3)^T\) has infinitely many solutions. (You don't need to solve the system in this part.) (e) Find all values of \(k\) for which the system \(Bx = (10, 1, 2014)^T\) has exactly one solution. (You don't need to solve the system in this part.)
Solution
Let's put the matrix \(B = \begin{bmatrix} 1 & 1 & 1 \\ 2 & 4 & 4 \\ 3 & 7 & k \end{bmatrix}\) into LU form. We get \(\begin{bmatrix} 1 & 1 & 1 \\ 0 & 2 & 2 \\ 0 & 4 & k - 3 \end{bmatrix}\) by subtracting 2 lots of row 1 from row 2 and 3 lots of row 1 from row 3. We get \[ U = \begin{bmatrix} 1 & 1 & 1 \\ 0 & 2 & 2 \\ 0 & 0 & k - 7 \end{bmatrix} \] by subtracting 2 lots of row 2 from row 3. We see \[ L = \begin{bmatrix} 1 & 0 & 0 \\ 2 & 1 & 0 \\ 3 & 2 & 1 \end{bmatrix}. \] (a) By setting \(k = 10\) we obtain \(A = LU\) where \(L = \begin{bmatrix} 1 & 0 & 0 \\ 2 & 1 & 0 \\ 3 & 2 & 1 \end{bmatrix}\) and \(U = \begin{bmatrix} 1 & 1 & 1 \\ 0 & 2 & 2 \\ 0 & 0 & 3 \end{bmatrix}\). (b) We spot that \((1, 1, 1)^T\) is a solution. \(L\) and \(U\) are invertible since their diagonals are nonzero, and so \(A\) is invertible. Thus this is the only solution. Alternatively, we can solve \(Lc = (3, 10, 20)^T\) via substitution to give \(c = (3, 4, 3)^T\), and \(Ux = c\) via substitution to give \(x = (1, 1, 1)^T\). (c) \(B\) is singular if and only if \(U\) is singular, and \(U\) is singular if and only if the last row is zero, i.e. if \(k = 7\). (d) This equation always has a solution since the first column is \((1, 2, 3)^T\). When \(k \neq 7\) the matrix is invertible and there is only one solution. When \(k = 7\) the vector \((0, 1, -1)^T\) is in the nullspace, so there are infinitely many solutions: in fact, the solutions are \((1, 0, 0)^T + c(0, 1, -1)^T\) for \(c \in R\). (e) When \(k \neq 7\) the matrix \(B\) is invertible, so there exists exactly one solution.Keywords: LU factorization, singular matrix, invertibility, unique solution