EXAM 1

Math 104 Linear Algebra with Applications, Section 2 Instructor: Guangliang Chen Duke University, Spring 2012 You have 75 minutes. No notes, no books, no calculators. You must show ALL work and explain your reasoning CLEARLY to receive full credit. Good luck! Name: ID number: SOLUTIONS

1. 2. 3. Signature: 4. 5. 6. 7. Total Score: 8. (/100 points) Date: I have adhered to the Duke Community Standard in completing this exam.

1. (18 pts) True or False (you do not need to justify your answer). (a) (x + y) (x y) = x2 y2 for any vectors x, y Rn . True. (b) Two unit vectors x, y Rn are orthogonal if and only if x + y = x y. True. x + y2 = 2 + 2x y while x y2 = 2 2x y. (c) There exists a 2 2 matrix such that rankA = 2 and rank (A2 ) = 1. False. If rankA = 2, then A is nonsingular, and so is A2 . (d) There exists a 2 2 matrix such that rankA = 1 and rank (A2 ) = 0. (A matrix has rank zero if and only if it is a zero matrix.) ( ) 0 1 True. A = is such an example. 0 0 (e) There exists an (n + 1) n matrix A such that for every vector b Rn+1 the equation Ax = b is consistent. False. If A is an (n + 1) n matrix, then rankA is at most n and can never achieve n + 1. Accordingly, the equation can be inconsistent for some b. (f) There exists an m n matrix A such that for some b Rm the equation Ax = b is inconsistent and for other b Rm the equation Ax = b always has a unique solution. True. Any m n matrix with rankA = n < m is such an example. (g) A Mmn has a left inverse if and only if rankA = m. False. Left inverse if and only if rankA = n. (h) Any square matrix can be written as a product of elementary matrices. False. Any invertible matrix can be written as such a product. (i) For any m n matrix A, the two equations AT Ax = 0 and Ax = 0 always have the same solution. True. In homework 2.5.15, we know that any solution to AT Ax = 0 is also a solution to Ax = 0. The other direction is obvious.

 1 2 1 2. (10 pts) Let A = 2 1 8. 3 5 5 (a) Find an invertible matrix E such that EA = U is in reduced REF. Solution. Start by forming the augmented matrix (A|I3 ) and apply elementary row operations to obtain RREF 1 2 1 | 1 1 0 5 | 5 0 2 1 0 1 2 | 3 0 1 = (U|E). (A|I3 ) = 2 1 8 | 3 5 5 | 1 0 0 0 | 7 1 3

(b) Solve the homogeneous equation Ax = 0. Solution. If we form the augmented matrix (A|0) and perform the same row operations as above, we must obtain (U|0). The corresponding equations are x1 + 5x3 = 0 x2 + 2x3 = 0 Clearly, the general solution is (x1 , x2 , x3 ) = (5, 2, 1)t, where t is any real number.

3. (10 pts) Let

( ) 2 1 9 A= . 1 3 2

Determine whether the matrix has a left or right inverse, and then nd one such inverse. Solution. It is very easy to check that rankA = 2. Thus, A has a right inverse. To nd one such inverse, we need to form the augmented matrix (A|I2 ) and apply elementary row operations to obtain RREF for A ( ) ( ) 1 2 1 9 | 1 1 0 5 | 3 5 5 . 1 3 2 | 1 0 1 1 | 1 2 5 5 3 1 5 5 The matrix 1 2 is a right inverse. 5 5 0 0

4. (20 pts) (a) Find parametric equations for the line that passes through a point x0 = (1, 2, 0) and is parallel to another line L : x = (1 + s, s, 3 2s). Solution. x = (1, 2, 0) + s(1, 1, 2) = (1 + s, 2 s, 2s)

(b) Find parametric equations for the plane P that contains the line L in part (a) and the origin. Solution. The line L contains a point (1, 0, 3) and a vector (1, 1, 2). We know that the origin is also in the plane. This gives a second vector in the plane (0, 0, 0) (1, 0, 3) = (1, 0, 3). Thus, the plane P has the following parametric equation: x = (1, 0, 3) + s(1, 1, 2) + t(1, 0, 3).

(c) Find a normal vector to the plane P in part (b). Solution. Only need to solve the following system: ( ) 1 1 2 | 0 , 1 0 3 | 0 that is, a1 a2 2a3 = 0 a1 3a3 = 0 One solution is (a1 , a2 , a3 ) = (3, 5, 1), which is normal to the plane.

(d) What is the distance from the point (1, 1, 0) to the plane P in part (b)? Solution. Using (1, 0, 3) as a referece point in the plane P, the distance from the point (1, 1, 0) to the plane P is Proja ((1, 1, 0) (1, 0, 3)) = a (0, 1, 3) 8 8 8 a = a = a = . 2 a 35 35 35

Alternatively, since P also contains the origin, we may have used the origin as the reference point. In this case, the distance from the point (1, 1, 0) to the plane P is Proja (1, 1, 0) = a (1, 1, 0) 8 8 8 a = a = a = . 2 a 35 35 35

5. (10 pts) Suppose A is a square matrix satisfying A2 = O. Answer the following questions. (If your answer is yes, nd the inverse; otherwise explain why an inverse cannot exist.) (a) Is A invertible? No. If A is invertible, so is AA = A2 = O, a contradiction.

The following argument also works: Suppose A1 exists, then A1 A2 = A1 O. That is, A = O. However, we know that the zero matrix does not have any inverse. Again, a contradiction.

(b) Is A 2I invertible? Yes. From (A 2I)(A + 2I) = A2 4I = O 4I = 4I, we obtain 1 (A 2I)1 = 4 (A + 2I).

6. (12 pts) Find the linear transformation T (x) in the plane that rst reects the point x about the line x2 2x1 = 0 and then rotates it 45 degrees counter-clockwise. Solution. Choose a point a = (1, 2)T in the line. The matrix corresponding to the reection is ( ) ( ) ( 3 4) ) 2 2 1 ( 1 0 5 5 T 1 2 aa I2 = 2 = . 4 3 2 2 2 0 1 a 1 +2 5 5 The matrix corresponding to the rotation is ) ( ) ( 2 cos sin 22 4 4 2 = 2 2 . sin cos 4 4
2 2

Therefore, the linear transformation is ( )( 2 22 3 2 5 T (x) = 2 2 4

4 5 3 5

) x.

7. (10 pts) Let A, B Mnn be two nonzero matrices whose product is zero: AB = O. (a) Show that any vector v Span(b1 , . . . , bn ) is a solution to the homogeneous equation Ax = 0. Proof. Since O = AB = A(b1 , . . . , bn ) = (Ab1 , . . . , Abn ), we have Ab1 = = Abn = 0. Thus, for any vector v = c1 b1 + + cn bn in the span, Av = A (c1 b1 + + cn bn ) = c1 Ab1 + + cn Abn = c1 0 + + cn 0 = 0. This shows that v is a solution to the homogeneous equation Ax = 0.

(b) Since B is a nonzero matrix, there must be at least one nonzero entry in B, say Bij = 0. Show that the ith column vector of A is a linear combination of the other columns of A. Proof. Write O = AB = (a1 , . . . , an )B. The jth column of AB is 0 = a1 B1j + + ai Bij + + an Bnj . Solving for ai , we get ai = B1j Bi1,j Bi+1,j Bnj a1 ai1 ai+1 an , Bij Bij Bij Bij

which requires Bij = 0.

8. (10 pts) A square matrix A is said to be orthogonal if its columns are all unit vectors and mutually orthogonal. That is, ai ai = 1 and ai aj = 0 for i = j. Prove that (a) Any orthogonal matrix A Mnn is invertible, and the inverse is its transpose. Proof. Let A = (a1 , . . . , an ) be an orthogonal matrix. Then the (i, j) entry of AT A is ( ) ( T ) A A ij = AT i aj = ai aj = 1 if i = j, and 0, otherwise. This indicates that AT A = In . Since A is square, we know that AT is the inverse of A.

Note that we cannot obtain AAT = In directly for orthogonal matrices. The reason is that T a1 . T AA = (a1 , . . . , an ) . = a1 aT + + an aT , . 1 n aT n a sum of n n matrices!

(b) For any orthogonal matrix A Mnn and any vector x Rn , we have Ax = x. (Hint: square both sides and show that they are equal.) Proof. Method 1: Ax2 = Ax Ax = x AT Ax = x In x = x x = x2 . Method 2: Ax2 = (Ax)T Ax = xT AT Ax = xT In x = xT x = x2 .