Math 415 Exam I. Name: Student ID: Calculators, books and notes are not allowed!

Transcription

1 Math 415 Exam I Calculators, books and notes are not allowed! Name: Student ID:

2 Score: Math 415 Exam I

3 (20pts) 1. Let A be a square matrix satisfying A 2 = 2A. Find the determinant of A. Sol. From A 2 = 2A, we get (deta)(deta) = det(2a) = 2 n deta, where n is the number of columns in A. Thus deta = 0 or deta = 2 n.

4 (20pts) 2. Let M 3 3 be the vector space of all real matrices of size 3 3, equipped with the matrix addition and scalar multiplication. Let W be the set of all 3 3 upper triangular real matrices. Prove that W is a subspace of M 3 3. Proof. We only need to prove that W is closed under the matrix addition and scalar multiplication. a 11 a 12 a 13 b 11 b 12 b 13 First, let 0 a 22 a 23, 0 b 22 b 23 be any two matrices in W. From 0 0 a b 33 the definition of the matrix addition, a 11 a 12 a 13 0 a 22 a a 33 + b 11 b 12 b 13 0 b 22 b b 33 which is an upper triangular matrix in W. = a 11 + b 11 a 12 + b 12 a 13 + b 13 0 a 22 + b 22 a 23 + b a 33 + b 33 Second, let c be any real number (a scalar). From the definition of the scalar multiplication, a 11 a 12 a 13 ca 11 ca 12 ca 13 c 0 a 22 a 23 = 0 ca 22 ca 23, 0 0 a ca 33 which is an upper triangular matrix in W. Therefore W is closed under the matrix addition and scalar multiplication and we prove that W is a subspace of M 3 3.,

5 (20pts) 3. Find the LU factorization of A = Sol. Perform the following elementary row operations for A: 1) Add ( 2)row 1 to row 2, 2) Add ( 1)row 2 to row 3. By operations 1) and 2), A can be reduced to U = Let L 1 be the elementary matrix associated to operation 1). And Let L 2 be the elementary matrix associated to operation 2). Then L 1 1 = 2 1 0, and L 1 2 = Thus A = L 1 1 L 1 2 U = which yields the LU factorization of A ,

6 (25pts)4. Let A = a) Find a basis for RangeA. b) Find a basis for KerA c) Find the general solution to the linear system A x =. 0 b b, where b is a real number. Sol. a) Perform the following elementary row operations for A: 1) Add ( 2)row 1 to row 2, 2) Add ( 1)row 2 to row By operations 1) and 2), A can be reduced to U = Since the pivots 1 and 4 are in the first column and the second column respectively, the first column and the second column in A are linearly independent. Therefore a basis of RangeA is {(1, 2, 0) T, ( 1, 2, 4) T }. b) To obtain KerA, we need to solve the homogeneous linear system U x = 0. This homogeneous liner system can be written as { x1 x 2 + x 4 = 0 4x 2 + 3x 3 3x 4 = 0 Then we get x 1 = 3x 4 3 1x 4 4 and x 2 = 3x x 4 4 for fixed x 3, x 4. Thus x 1 3 x = x 2 x 3 = x 4 3 1x 4 4 3/4 1/4 3 4 x x 4 4 x 3 = x 3/ x 3/4 4 0 x 4 x The fundamental theorem of Linear Algebra yields dim(kera) = 4 dim(rangea) = 4 2 = 2. Thus a basis of Ker(A) is {( 3/4, 3/4, 1, 0) T, (1/4, 3/4, 0, 1) T }. c) Since we know KerA is span{( 3/4, 3/4, 1, 0) T, (1/4, 3/4, 0, 1) T } from part b), we only need to find a particular solution to A x = (0, b, b) T in order to obtain the general solution. By performing the operations 1) and 2) in part a), we reduce the augmented matrix b b to b The corresponding linear system is { x1 x 2 + x 4 = 0 4x 2 + 3x 3 3x 4 = b Clearly x = (b/4, b/4, 0, 0) T is a particular solution to this linear system. Thus the general solution is x = (x 1, x 2, x 3, x 4 ) T = (b/4, b/4, 0, 0) T + c 1 ( 3/4, 3/4, 1, 0) T + c 2 (1/4, 3/4, 0, 1) T.

7 (15pts) 5. Let A be an invertible matrix of n n. And let v 1, v 2,, v k be (column) vectors in R n. a) Prove that if v 1, v 2,, v k are linearly independent, then A v 1, A v 2,, A v k are linearly independent. b) Let M 1 = ( v 1 v k ) (the matrix generated by v1, v 2,, v k ), and M 2 = ( A v 1 A v k ) (the matrix generated by A v 1, A v 2,, A v k ). Prove that rankm 1 = rankm 2. Proof. Proof of Part a). Let c 1 A v 1 + c 2 A v 2 + c k A v k = 0, where c 1, c 2,, c k are scalars. To prove the linear independence of A v 1, A v 2,, A v k, we only need to prove c 1 = c 2 = = c k = 0. Notice that Thus we obtain c 1 A v 1 + c 2 A v 2 + c k A v k = A(c 1 v 1 + c 2 v c k v k ). A(c 1 v 1 + c 2 v c k v k ) = 0. (1) Since A is invertible, A 1 exists. Multiply both sides of (1) by A 1 to get which yields A 1 A(c 1 v 1 + c 2 v c k v k ) = A 1 0, c 1 v 1 + c 2 v c k v k = 0. Since in part a) v 1, v 2,, v k are linearly independent, we get that c 1 = c 2 = = c k = 0 from the definition of linear independence. Thus the proof of part a) is completed. Proof of Part b). Method 1. From Part a), one can conclude that v 1, v 2,, v k are linearly independent if and only if A v 1, A v 2,, A v k are linearly independent. This is because v j = A 1 (A v j ) for all j. Thus we can conclude rankm 1 = rankm 2 from the definition of the rank. Method 2. Set r 1 = rankm 1 and r 2 = rankm 2. First, we prove r 1 r 2. From the definition of the rank, we know there are r 1 linearly independent vectors among v 1, v 2,, v k. Let us name these r 1 linearly independent vectors by u 1,, u r1 respectively. From Part a), A u 1,, Au r1 are linearly independent. Thus we get r 1 linearly independent vectors among A v 1, A v 2,, A v k. Thus the maximal number of linearly independent vectors among A v 1, A v 2,, A v k must at least be r 1, which gives r 1 r 2 by the definition of the rank. Now we prove r 2 r 1. This can be done similarly. In fact, let w 1 = Av 1, w 2 = Av 2,, w k = Av k. Since A is invertible, we have v 1 = A 1 w 1, v 2 = A 1 w 2,

8 , v k = A 1 w k. Repeat the same argument as in the previous paragraph to conclude rank( w 1 w k ) rank(a 1 w 1 A 1 w k ), which gives r 2 r 1 since ( w 1 w k ) = M 2 and (A 1 w 1 A 1 w k ) = M 1. Finally we conclude r 1 = r 2 since r 1 r 2 and r 2 r 1. And this completes the proof.