Page 1 Lab 5 Elementary Matrix and Linear Algebra Name ____________________ Due at Final Exam Spring 2011 Score = ____/25 1. (3 points) Go to AKiTi Miscellaneous Mathematical Utilities page at http://www.akiti.ca/Mathfxns.html and numerically determine the eigenvalues and eigenvectors of the following matrices. Fill in the table and printout and attach the results from the program. a) 2 3 1 A 5 7 9 7 4 8 det A 1 2 3 1 2 3 b) 1 2 4 3 3 8 0 1 A 1 0 6 2.5 2 1 2.5 2 det A 1 2 3 4 1 2 3 4 2. (3 points) a) Prove that the transpose of a matrix has the same eigenvalues as the original matrix. Is the same true of the eigenvectors? Explain your answer. b) For real symmetric matrices the eigenvectors can always be expressed using only real numbers. Is this true of any real square matrix? Explain your answer. 3/9/2016 Madison Area Technical College Page 2 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 3. (11 points) Complex Inner Product Spaces, Unitary and Hermitian matrices. The concept of an inner product can be extended to vectors consisting of complex entries. In analogy with R n , the space of n dimensional vectors with complex components is called n . The inner or Hermitian product of two vectors in is defined as u | v u, v uT v n n u v j 1 j j . Here z designates the complex conjugate of z. Recall that for a complex number z, z a bi with a and b real, the complex conjugate is defined as z a bi . Furthermore, z zz a bi a bi a 2 b 2 0 . The convention of complex conjugation “on the right” in 2 the definition of u | v is arbitrary and some authors define the Hermitian product with complex conjugation on 5 2i the left. What’s important is to be consistent. As an example of the Hermitian product, let u 3 and 4i 3 7i v 2i , then u | v 5 2i 3 7i 3 2i 4i 1 i 5 31i . 1 i a) Show that the Hermitian inner product satisfies the following properties. Here u, v, and w represent vectors and c is a complex scalar. _______ u|v v|u uv| w u| w v| w w| u v w| u w| v cu | v c u | v u | cv c u | v u | u 0 with equality if and only if u 0 . The Cauchy Schwartz inequality is still true. For any complex scalar and non-zero vectors u and v , _______ u v | u v 0 . So, u v | u v u | u v | u v | u then u | u 2 v|u v|v u|u v|v v|u 3/9/2016 2 v|u v|v 2 2 2 v|v u |u v|u v|v 2 v | v 0 . Let 2 0 . Rearranging this inequality gives with equality if and only if u cv for some complex scalar c. Madison Area Technical College u|v v|v , Page 3 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 The adjoint or Hermitian conjugate of an n x n complex matrix is defined as A† AT , i.e., the transpose of the matrix of complex conjugates. Note: when moving a matrix multiplier from left to right in an inner product one transforms to the Hermitian conjugate. Au | v Au v u A v u T T T T _______ T A v u | A†v . An n x n complex matrix U is called unitary if and only if U † U 1 , i.e., U †U UU † I n . An n x n complex matrix A is called Hermitian if and only if A† A . b) Which of the following matrices are unitary? 1 2 1 A 2 0 0 3 i 2 2 0 3 i 2 2 0 0 0 1 i 2 i 2 1 0 2 1 0 B 2 1 i 0 2 i 0 2 3 i 2 2 0 3 i 2 2 0 0 0 1 i 2 i 2 0 0 C 1 i 2 i 2 1 2 1 2 0 0 1 i 2 1 i 2 0 1 3 i 2 2 0 0 1 3 i 2 2 0 0 0 i 2 i 2 sin cos sin i cos cos D sin sin cos i cos sin i cos 0 sin c) What is true about the determinant of a unitary matrix? d) Is every orthogonal matrix a unitary matrix? Explain your answer. e) Look up the definition of a group in algebra. Show that the set of n x n unitary matrices form a group with respect to the operation of matrix multiplication. f) Let SU(n) be the set of all n x n unitary matrices with a determinant of one. Show that SU(n) is a group with respect to the operation of matrix multiplication. Find an element of SU(2) that is not the identity matrix. 3/9/2016 Madison Area Technical College Page 4 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 g) Suppose A is an n x n matrix and U is an n x n unitary matrix. What is the det U † AU ? What is the trace U † AU ? h) Can a non-identity matrix be both Hermitian and unitary? Explain your answer. i) For any matrix below which is Hermitian determine its eigenvalues. 3 5 5 11 3 5 5 11 3i 5 5 11 3i 5 5 11i 3 5i 5i 11 3 5i 5i 11 __________ j) Suppose A is n x n Hermitian matrix and u is a vector in n . What is Au | u Au | u ? k) Suppose A is n x n Hermitian matrix and for a non-zero , A . What is ? Choose one of the following six problems. Each problem is worth 8 points. You may do extra problems to earn up to 40 extra lab points. 4. The Spectral or Fourier Representation of Hermitian Matrices a) If A is an n x n Hermitian matrix then it has n linearly independent eigenvectors which span n j j1 n . Let designate a set of n normalized linearly independent eigenvectors of A with A j j j and j | j 1. Determine j | k . b) Let B be an n x n matrix defined as Bi j i ' th component of j , i.e., B 1 2 n . † Determine BB . 3/9/2016 Madison Area Technical College Page 5 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 Wolfram Alpha can be used to determine the eigenvalues and a set of eigenvectors for a matrix as illustrated below. c) Determine the eigenvalues and a set of orthonormal eigenvectors of the matrix A. 5 A 3 i 3 i 2 1 ________________ 1 ________________ 2 ________________ 2 ________________ 2 . Compute the following: i Let v 2 j 1 j v | j j ________________ 3/9/2016 Av ________________ Madison Area Technical College Page 6 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 d) Determine the eigenvalues and a set of orthonormal eigenvectors of the matrix A. 11 8 3 A 8 3i 4 3 8 9 8 3i 4 3i 4 3 i 4 5 2 1 ________________ 1 ________________ 2 ________________ 2 ________________ 3 ________________ 3 ________________ 2i Let v i . Compute the following: 1 3 j 1 j v | j j ________________ Av ________________ e) Let A be an n x n Hermitian matrix with an orthonormal set of eigenvectors j | k jk . Let v be any vector in n What does the sum j 1 j n n j j1 so that A j j j and . v | j j represent? n If A is non-singular, what does the sum 1 j 1 3/9/2016 v | j j represent? j Madison Area Technical College Page 7 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 Problems 5 and 6 are dedicated to the memory of Professor C. H. Blanchard (1923 – 2009) of the University of Wisconsin-Madison Physics Department. He was the best and most inspiring teacher I have ever encountered. 5. The Moment of Inertia Tensor In the analysis of the rotation of a rigid body one defines a second rank tensor, called the moment of inertia tensor which can be represented by the following 3 x 3 real symmetric matrix. I inertia I xx I xy I xz I xy I yy I yz I xz I yz I zz For a mass density m x, y, z , the matrix elements of the moment of inertia tensor are defined by the following triple integrals over the region of three dimensional space S bounded by the surface of the rigid body. I xx x, y, z y 2 m z 2 dV S m x, y, z xydV I xy S m x, y, z xzdV I xz S I yy x, y, z x 2 m z 2 dV S m x, y, z yzdV I yz S I zz x, y, z x m 2 y 2 dV S x For an angular velocity vector y , the angular momentum of the rigid body is given by z 1 I inertia and the kinetic energy of rotation is T I inertia . For a uniform solid right circular cylinder of radius a 2 and height h inclined as shown in Figure 1, the moment of inertia tensor is I inertia 3/9/2016 h 2 5a 2 4 16 0 3 h2 a 2 4 3 4 0 h2 a 2 3 4 0 3 h2 a 2 4 3 4 0 2 2 h 7a 12 16 Madison Area Technical College Page 8 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 Figure 1 a) Determine the eigenvalues and an orthonormal set of eigenvectors for this I inertia . b) The orthonormal eigenvectors are called the principal axes of the rigid body. Explain this terminology. 3/9/2016 Madison Area Technical College Page 9 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 6. Coupled Pendulums A simple pendulum consists of an object of mass m suspended by an essentially weightless and non stretchable string of length that is allowed to swing freely. As in figure 2 let the angle denote the displacement from the downwards vertical. is related to the standard polar angle by 3 , so that the position vector of the mass 2 3 3 iˆ sin ˆj sin iˆ cos ˆj . The acceleration vector is thus is given by R cos 2 2 2 d ˆ ˆ d sin iˆ cos ˆj cos i sin j 2 2 dt dt dt 2 2 d 2 ˆ d 2 d d sin cos i cos sin dt dt dt 2 dt 2 a d 2R 2 ˆj Figure 2 3/9/2016 Madison Area Technical College Page 10 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 The forces acting on the mass are the tension in the string, T , which points to the origin and gravity which points straight down. F T sin iˆ cos ˆj mg ˆj T sin iˆ T cos mg ˆj . From Newton’s second law d of motion, F ma m sin dt 2 d 2 ˆ i m cos dt 2 d cos dt 2 d 2 ˆ j , so that we have sin dt 2 the coupled differential equations d T sin m sin dt d T cos mg m cos dt d From the first equation T m dt 2 cos d 2 sin dt 2 2 d cos 2 dt 2 2 d sin 2 dt 2 . Substituting this expression into the second equation gives the following. d m dt cos d cos sin dt 2 2 d mg m cos dt cos 2 d 2 d 2 g sin sin dt 2 dt 2 2 2 d 2 d 2 2 2 sin cos 2 2 dt dt 2 g sin d 2 dt 2 d sin 2 dt 2 g sin We seek solutions that satisfy the initial values of 0 and 0 . This non linear second order initial value problem has a solution which can be expressed as an elliptic function. A simpler result is obtained if we restrict our attention to small angles where sin . The differential equation then is linear and is given by 3/9/2016 Madison Area Technical College d 2 dt 2 g . Page 11 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 Looking for exponential solutions of the form t et gives the auxiliary equation 2 i g g which has solutions i0 . Thus, t b1ei0t b2ei0t b1 b2 cos 0t i b1 b2 sin 0t . Requiring that the _____ solution be real means that t t , so that b1 b2 cos 0t i b1 b2 sin 0t b1 b2 cos 0t i b1 b2 sin 0t . Since sine and cosine are linearly independent functions it must be true that b1 b2 b1 b2 . b1 b2 b2 b1 c ic2 d id2 Taking b1 1 for real c1 , c2 , d1, d 2 yields d 2 c2 and d1 c1 so that b1 b2 c1 and , b2 1 2 2 b1 b2 ic2 . So the general solution in the small angle approximation is t c1 cos 0t c2 sin 0t . Fitting the initial conditions gives t 0 cos 0t T 2 0 2 g and a frequency of f 0 sin 0t . The solution is periodic with a period given by 0 g 1 . If the pendulum is “released from rest”, 0 0 and T 2 t 0 cos 0t , so that 0 is the amplitude of the angle oscillations. Two pendulums each of length are coupled by a cross string as shown on Figure 3. The effect of the cross string can be modeled by a force on pendulum 1 by pendulum 2 equal to k 1 2 where k is a positive proportionality constant. By Newton’s third law of motion the force on pendulum 2 by pendulum 1 is k 2 1 . In the small angle approximation the two angles satisfy the following linear differential equation. g k d 2 1 dt 2 2 k 3/9/2016 1 g k 2 k Madison Area Technical College Page 12 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 Figure 3 Graphic from page 17 of Nine Experiments for a Third Grade Hour, 2009 Memorial Edition by C. H. Blanchard. Figure 4 3/9/2016 Madison Area Technical College Page 13 g k Let A k Lab 5 Elementary Matrix and Linear Algebra Spring 2011 2 . A is a real symmetric matrix. It has orthonormal eigenvectors 1 , 2 that span R . g k k 2 d 2c 2 1 t d 2 1 d c1 c t c t So for any value of t, 2 A c11 c 2 2 . 1 1 2 2 . Thus, 2 2 1 2 t dt dt dt 2 2 a) Display the second order differential equations for both c1 t and c2 t . b) Solve for both c1 t and c2 t with initial conditions for each defined by the values of 1 0 , 2 0 , 1 0 , and 2 0 . c) Use the solutions for c1 t and c2 t to display the solutions for 1 t and 2 t . d) Is the motion of the coupled pendulums always periodic, i.e. is there a period T with both 1 t T 1 t and 2 t T 2 t for all values of t? Explain your answer. What conditions on the constants k , , g is required for periodicity? 3/9/2016 Madison Area Technical College Page 14 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 e) Display the solutions for 1 t and 2 t if both masses are released from rest, i.e. 1 0 2 0 0 . f) If 1 0 2 0 0 and 2 0 0 describe the motion. In particular, does the second pendulum oscillate? g) Display the solutions for 1 t and 2 t if1 0 2 0 0 and 2 0 1 0 0 . h) Display the solutions for 1 t and 2 t if1 0 2 0 0 and 2 0 1 0 0 . i) The solutions of g) and h) are periodic solutions which consist of oscillations at a single frequency. These are called the normal modes of the system. One of these modes is called a “walking mode”. Which one is it? j) One of the normal modes is faster, i.e. has a higher frequency. Which normal mode is it? k) What is the connection between the eigenvectors of A and the normal modes? l) What is the connection between the eigenvalues of A and the normal modes? 3/9/2016 Madison Area Technical College Page 15 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 7. Rayleigh-Ritz Variational Calculations of Extreme Eigenvalues Suppose A is an n x n real symmetric matrix. In many applications such as bound states in Quantum Mechanics and economic optimizations the extreme eigenvalues max and min are of particular interest. Consider an n x 1 vector called a “trial function”. The eigenvectors of A are linearly independent and an orthonormal basis for R n consisting of eigenvectors of A is always possible. Designate this orthonormal basis by j A j j j , j , k jk , and max 1 2 n j1 with n min . The trial function can be expanded as , j j . n j1 n n n j 1 k 1 j 1 k 1 A, A , j j , , k k , j , k A j , k n j , j , k j , k j , j , k jk j , j n n n j 1 k 1 n n j 1 k 1 j 1 , A, For 0 , the Rayleigh Quotient is defined as n , min , j n j 1 , , j 2 j , . It can be bracketed as follows: , A , n , n min 2 j 1 j 1 max 2 j , , n 2 2 2 A, j 1 , , max , A, min , max j 1 j j Any choice of 0 gives a lower bound to max and an upper bound to min . By allowing to vary over all non zero vectors of R n we have the following characterizations. A, R , 0 , A, max max R , 0 , min min n n 3/9/2016 Madison Area Technical College Page 16 Lab 5 Elementary Matrix and Linear Algebra If varies over just a subspace W of R n then min A, or if 2 W W , 0 , then max max A, W , 0 , max and A, W , 0 , min A, max A, and max . If 1 W W , 0 , W , 0 , min then min A, . How well the variational calculations W , 0 , min approximate the extreme eigenvalues depends on “how close” the subspace W is to 1 and 2 .Let Spring 2011 , , then A, A, with , 1 . So in practice , variational calculations are done using normalized trial functions, i.e. , 1. This is equivalent to varying the trial function over a subspace of the unit n-sphere. As an example consider the 2 x 2 real symmetric matrix a c A . The eigenvalues are solutions of the characteristic polynomial 2 a b ab c2 0 . c b max 1 min 2 ab a b 2 4c 2 2 ab a b 2 4c 2 2 ab a b 2 c 2 2 ab a b 2 c 2 2 2 2 cos with 0 2 spans the unit circle (the unit 2-sphere). sin The trial function Thus, 1 max A, and 2 min 0 2 0 2 A, . This can be verified as follows. a cos c sin cos 2 2 f A , , a cos 2c cos sin b sin c cos b sin sin Using the trigonometric identities cos 2 3/9/2016 1 cos 2 2 ; sin 2 1 cos 2 2 ; sin cos sin 2 2 Madison Area Technical College Page 17 f Lab 5 Elementary Matrix and Linear Algebra a b a b cos 2 c sin 2 ; cos 2 2 a b a b 2 2 c 2 2 Spring 2011 ; sin c a b 2 c 2 2 ab a b 2 c cos cos 2 sin sin 2 2 2 2 2 ab a b 2 c cos 2 2 2 ab a b 2 So the maximum of f at is c max and the minimum of f at is 2 4 2 2 2 2 ab a b 2 c min . 2 2 2 cos For a 3 x 3 real symmetric matrix, trial functions of the form 1 sin would in general give only 0 approximations to max and min since 1 does not span all of the unit 3-sphere. The trial function sin cos 2 sin sin with 0 2 ; 0 does span the unit 3-sphere and so cos max max A 2 , 2 and min min A 2 , 2 . 0 2 0 0 2 0 sin sin cos sin sin sin For a 4 x 4 real symmetric matrix, trial functions of the form with sin cos cos 0 2 ; 0 ; 0 would give the exact values of max and min , but this would involve finding extreme values of a function of three independent variables. 3/9/2016 Madison Area Technical College Page 18 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 Fill in the following tables for each matrix below. a) 10 6 A 6 5 cos sin min f min max f max f A, b) 7 4 A 1 2 cos sin min f min max f max f A, c) 5 A 3 i 3/9/2016 3 i 2 cos sin min f min max f max f A, Madison Area Technical College Page 19 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 d) 4 0 6 A 0 2 0 6 0 4 cos sin 0 min f min max f max f A, e) 4 0 6 A 0 2 0 6 0 4 sin cos sin sin f , A, cos min f min max f max f) 11 0 A 0 5 3/9/2016 5 0 0 0 0 11 0 0 4 6 6 4 cos 0 sin 0 min f min max f max f A, Madison Area Technical College Page 20 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 g) 11 0 A 0 5 sin cos 0 f , A, sin sin cos 5 0 0 0 0 11 0 0 4 6 6 4 min f min max f max h) For some of the above cases the interval min , max was not identical to f min , f max . Explain why in each case. i) (2 point bonus) The Rayleigh-Ritz variational method can be used to find the extreme eigenvalues of complex Hermitian matrices. Explain how the trial functions need to be chosen to make this work, 8. Iterative Scheme to Calculate Eigenvalue of Greatest Absolute Value Consider a real n x n matrix A with n distinct real eigenvalues. Let be the most extreme eigenvalue (i.e., the eigenvalue of largest absolute value) of A with eigenvector 1 and designate the remaining eigenvectors of A as k k 2 with A k k k . Let be a vector in R n . Because the eigenvectors of A form a basis for R n , has n an expansion in the eigenspace of A: c11 Then Am , Am 3/9/2016 n n c , c k 1 k k j 1 j n c k 2 k k . Assume that ck 0 . n n n m c c , kmck c j k , j . j 1 j 1 j j 1 k 2 j 1 Madison Area Technical College Page 21 Lab 5 Elementary Matrix and Linear Algebra A A , m 1 m , m 1c1 c j 1 , j km 1ck c j k , j n n n j 1 n k 2 n n j 1 k 2 j 1 j 1 m c1 c j 1 , j km ck c j k , j n n c1 c j 1 , j k k 2 j 1 n n m c1 c j 1 , j k k 2 j 1 m 1 c1 c j 1 , j k j 1 k 2 n n n m 1 n ck c j k , j j 1 m n ck c j k , j j 1 m 1 ck c j k , j n j 1 c1 c j 1 , j k ck c j k , j j 1 k 2 j 1 n m n n k 1 , all of the sums ck c j k , j and Since k 2 j 1 k Spring 2011 m n k k 2 n m 1 ck c j k , j will approach zero n j 1 A , converges to . Fill in the following tables. as m increases. Thus as m increases the ratio A , m 1 m 10 6 A 6 5 a) ______________ 1 1 A , A , m 1 m m 1 2 3 4 5 6 7 3/9/2016 Madison Area Technical College Page 22 Lab 5 Elementary Matrix and Linear Algebra 7 4 A 1 2 b) ______________ Spring 2011 2 1 A , A , m 1 m m 1 2 3 4 5 6 7 7 4 A 1 2 c) ______________ 1 1 A , A , m 1 m m 1 2 3 4 5 6 7 3/9/2016 Madison Area Technical College Page 23 Lab 5 Elementary Matrix and Linear Algebra 4 0 6 A 0 2 0 6 0 4 d) ______________ A , A , Spring 2011 1 1 1 m 1 m m 1 2 3 4 5 6 7 11 0 A 0 5 e) 5 0 0 0 0 11 0 0 4 6 6 4 ______________ A , A , 1 1 1 1 m 1 m m 1 2 3 4 5 6 7 f) Why are the results for b) and c) different? 3/9/2016 Madison Area Technical College Page 24 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 9. Eigenvalues of a Tridiagonal Toeplitz Matrix Consider the n by n matrix G(n), defined by the following formula: G n i , j ai , j bi 1, j ci 1, j 1 i, j n a c 0 0 i.e., G n 0 0 0 b a c 0 0 0 0 0 b a c 0 0 0 0 0 b a 0 0 0 0 0 0 0 a c 0 , 0 0 0 0 b a c 0 0 0 0 0 b a Matrices such as G(n) whose values along a given diagonal are constant are referred to as Toeplitz matrices. If bc = 0 the only eigenvalue of G(n) is a. To get a more useful result we must assume that bc 0 . Let X be an n component eigenvector of G(n) with eigenvalue . The components of X satisfy the equation bX k 1 a X k cX k 1 0 1 k n (1) with the boundary conditions that X 0 X n1 0 . (2) Equation (1) is a linear homogeneous recurrence relation of degree 2 and it can be solved in a manner analogous a c with solving a linear second order differential equation. Since X 0 0 and X k 1 X k 1 Xk , a b b nontrivial (i.e., nonzero) eigenvector requires that X1 be nonzero. Without loss of generality we can take X 1 1 . For a given eigenvalue this initial condition uniquely determines X. To see this, suppose Y is a second eigenvector with Y1 = 1. Because equation (1) is linear and homogeneous, Z = Y – X also satisfies equation (1). However, Z0 = Z1 = 0 so that the recursion requires that Z is the zero vector. 3/9/2016 Madison Area Technical College Page 25 Lab 5 Elementary Matrix and Linear Algebra Spring 2011 To construct the general solution of equation (1) we look for exponential solutions of the form X k r k with r 0 . 2 a c a and Hence, r k 1 br 2 a r c 0 which gives the two solutions r1 2b b 2b r2 2 c a . If the discriminant is zero, r1 r2 and the general solution of equation (1) would be 2b b 2b a of the form X k r1k kr1k for constants and . However, the boundary conditions of equation (2) would force both constants to be zero. Thus, a nontrivial eigenvector requires that r1 r2 . For this case the general solution of equation (4) is given by X k r1 k r2 k . The initial conditions X 0 0 and X 1 1 are satisfied when 1 . Using the equations for r1 and r2 to solve for gives r1 r2 a c c br1 a br2 . (3) r1 r2 c c This results in the following quadratic equation for r1 , r1 2 r1 r2 0 , so that br b 2 c c r2 r2 br br2 c 2 . Since r r , it must be true that r1 r1 r2 . The boundary condition at 1 2 2 b k = n + 1 can be expressed as X n1 1 r n1 r2 n1 0 , so that for any integer j, r1 n1 r2 n1e i 2 j . This r1 r2 1 equation has the following n distinct solutions each of which corresponds to a different nontrivial eigenvector: r1 r2 e i 2 j n1 2 j 2 j j i c i n1 c i e for 1 j n . Hence, r1 2 r1 r2 e n1 e n1 , so that r1 , where the radical b b designates the principal square root of a complex number. From equation (3) the n eigenvalues of G(n) are given by 3/9/2016 Madison Area Technical College Page 26 Lab 5 Elementary Matrix and Linear Algebra j a c i e j n 1 c b Spring 2011 j c i 2c j b e n1 a cos . (4) b c n 1 b It needs to be noted that for complex c and b it is not necessarily true that 2c 2 bc . The result could be off by a c b 1 1 minus sign (recall that i i 1 1 1 1 1 i ). On the other hand, 1 n 1 j 2c 2c j 2c j , so that allowing the index j to a a cos cos cos n 1 n 1 n 1 c c c b b b j will indeed generate all the eigenvalues of G(n). run from 1 to n in the expression a 2 bc cos n 1 j a Compute and display the five eigenvalues and associated eigenvectors of the following matrix. i 0 0 8 4i 8 i 0 0 4i 8 i 0 4i 8 0 0 0 0 4i 3/9/2016 0 0 0 i 8 Madison Area Technical College