sol5 - CAAM 336 · DIFFERENTIAL EQUATIONS Problem Set 5 ·...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: CAAM 336 · DIFFERENTIAL EQUATIONS Problem Set 5 · Solutions Posted Thursday 23 September 2010. Due Wednesday 29 September 2010, 5pm. All of the problems on this set use the inner product ( u,v ) = Z 1 u ( x ) v ( x ) dx. 1. [30 points: 6 points each for (a) and (b); 9 points each for (c) and (d)] Consider the linear operator L b : C 2 b [0 , 1] → C [0 , 1] defined by L b u =- d 2 u dx 2 , where C 2 b [0 , 1] = n u ∈ C 2 [0 , 1] : du dx (0) = u (1) = 0 o . (a) Is L b symmetric? (b) What is the null space of L b ? That is, find all u ∈ C 2 b [0 , 1] such that L b u ( x ) = 0 for all x ∈ [0 , 1]. (c) Show that ( L b u,u ) ≥ 0 for all nonzero u ∈ C 2 b [0 , 1] and explain why this implies that λ ≥ 0 for all eigenvalues λ . (d) Find the eigenvalues and eigenfunctions of L b . Solution. (a) Yes, L b is symmetric . To prove this, we shall show that for any u,v ∈ C 2 b [0 , 1], we have ( L b u,v ) = ( u,L b v ). He we shall use primes to denote derivation with respect to x . Integrating by parts twice, we have ( L b u,v ) = Z 1- u 00 ( x ) v ( x ) dx =- [ u ( x ) v ( x )] 1 + Z 1 u ( x ) v ( x ) dx =- [ u ( x ) v ( x )] 1 + [ u ( x ) v ( x )] 1- Z 1 u ( x ) v 00 ( x ) dx. Since u,v ∈ C 2 b [0 , 1] we have u (0) = 0 and v (1) = 0, and hence the first term in square brackets must be zero. Again using u,v ∈ C 2 b [0 , 1] we have v (0) = 0 and u (1) = 0, and hence the second term in square brackets is also zero. It follows that ( L b u,v ) = Z 1 u ( x )(- v 00 ( x )) dx = ( u,L b v ) . (b) The general solution to the differential equation- d 2 u dx 2 = 0 has the form u ( x ) = A + Bx for constants A and B that are determined by the boundary conditions. In order for u to be in C 2 b [0 , 1], we must have u (0) = 0; since u ( x ) = B , we have B = 0. Now u ∈ C 2 b [0 , 1] also requires u (1) = 0, and since u (1) = A , we conclude that A = 0 too, meaning that u ( x ) = A + Bx = 0 for all x ∈ [0 , 1]. Thus, the only element of the null space is the zero function, that is, the null space is trivial, N ( L b ) = { } . (c) [ GRADERS : please count ( L b u,u ) ≥ 0 for 5 points, and count λ ≥ 0 for 4 points.] We wish to show that ( L b u,u ) ≥ 0 for all ∈ C 2 b [0 , 1]. Using the first integration by parts from part (a), we have ( L b u,u ) =- [ u ( x ) u ( x )] 1 + Z 1 u ( x ) u ( x ) dx = Z 1 ( u ( x )) 2 dx. Thus, ( L b u,u ) is the integral of a nonnegative function, so it is nonnegative. This statement implies that all eigenvalues are non-negative, since λ ( u,u ) = ( λu,u ) = ( L b u,u ) ≥ , and we know that ( u,u ) > 0 for all nonzero u due to the positivity of the inner product. (Though not asked for in the problem statement, one can show that u ( x ) 6 = 0 for some x ∈ [0 , 1]: otherwise u would be a constant, and the only constant that satisfies the boundary conditions is u ( x ) = 0, which is prohibited from consideration in the problem statement. Thus, we can actually say ( L b u,u ) > 0 for all nonzero u .) (d) Suppose that...
View Full Document

This note was uploaded on 01/21/2012 for the course CAAM 330 taught by Professor Tompson during the Fall '09 term at UVA.

Page1 / 12

sol5 - CAAM 336 · DIFFERENTIAL EQUATIONS Problem Set 5 ·...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online