Numerical Methods Qualification Exam Problems and Solutions (University of Maryland)/January 2009

From Wikibooks, open books for an open world
Jump to navigation Jump to search

Problem 1[edit | edit source]


Solution 1[edit | edit source]

Problem 2[edit | edit source]


Solution 2[edit | edit source]

Problem 3[edit | edit source]

Let with . Assume

Problem 3a[edit | edit source]


It is known that the symmetric matrix can be factored as



where the columns of are orthonormal eigenvectors of and is the diagonal matrix containing the corresponding eigenvalues. Using this as a starting point, derive the singular value decomposition of . That is show that there is a real orthogonal matrix and a matrix which is zero except for its diagonal entries such that

Solution 3a[edit | edit source]

We want to show



which is equivalent to


Decompose Lambda[edit | edit source]

Decompose into i.e.



We can assume since otherwise we could just rearrange the columns of .

Define U[edit | edit source]

Let where


Verify U orthogonal[edit | edit source]