This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Math 471, Fall 2006 Homework 7 Solutions Assigned: Friday, October 27, 2005. Due: Friday, November 3, 2005 . 1. (Positive Definite Matrices) . Do page 221, #9. Some hints: Establish the 2 2 case. Then reduce the general problem to this special case using the definition of positive definiteness. Proof. Define a vector x whose entries are zero except for a one in the i th position. By the definition of a positive definite matrix, < x T Ax = a ii . The second part is harder. First, we prove the assertion in the 2 2 case. Suppose that A = a 11 a 12 a 12 a 22 is positive definite. Since the eigenvalues of a positivedefinite matrix are positive, the determinant of A is positive, and < det A = a 11 a 22 a 2 12 . Rearrange this equation to establish the 2 2 case. Next, observe that we can reduce the general problem to this special case. Suppose that A is an n n positivedefinite matrix. Without loss of generality, let i = 1 and j = 2. By considering vectors x of the form x = [ x 1 x 2 ... 0] T , we see that x 1 x 2 T a 11 a 12 a 12 a 22 x 1 x 2 > . So this 2 2 submatrix is positive definite. Applying the result for this special case gives a 2 12 < a 11 a 22 . Since we can make the same argument for any i 6 = j , the result follows. 2. (Poisson BVP) Do page 741, #7. Mesh size h RMS Error RMS Error / h 2 1/4 0.00136076 0.02177211 1/6 0.00056606 0.02037815 1/8 0.00030714 0.01965715 1/10 0.00019220 0.01921973 1/12 0.00013144 0.01892701 1/14 0.00009550 0.01871766 1/16 0.00007250 0.01856061 RMS error is the 2 error over all the points in the mesh divided by the squareroot of the number of points. The RMS error divided by h 2 tends to a constant . 019, so we have secondorder convergence. Is this also true for the error?...
View Full
Document
 Winter '08
 LinZhi
 Numerical Analysis, Matrices

Click to edit the document details