QR Decomposition - Using For Solution To Linear Inverse Problems

Using For Solution To Linear Inverse Problems

Compared to the direct matrix inverse, inverse solutions using QR decomposition are more numerically stable as evidenced by their reduced condition numbers .

To solve the underdetermined linear problem where the matrix A has dimensions and rank, first find the QR factorization of the transpose of A:, where Q is an orthogonal matrix (i.e. ), and R has a special form: . Here is a square right triangular matrix, and the zero matrix has dimension . After some algebra, it can be shown that the solution to the inverse problem can be expressed as: 
x = Q
\begin{bmatrix} (R_1^T)^{-1}b \\ 0 \end{bmatrix}
where is found by Gaussian elimination.

To find a solution to the overdetermined problem which minimizes the norm, first find the QR factorization of A: . The solution can then be expressed as, where and are the same as before, but now is a projection matrix that maps a vector in into .

Read more about this topic:  QR Decomposition

Famous quotes containing the words solution, inverse and/or problems:

    “Any solution to a problem changes the problem.”
    —R.W. (Richard William)

    “The quality of moral behaviour varies in inverse ratio to the number of human beings involved.”
    —Aldous Huxley (1894–1963)

    “The mother’s and father’s attitudes toward the child correspond to the child’s own needs.... Mother has the function of making him secure in life, father has the function of teaching him, guiding him to cope with those problems with which the particular society the child has been born into confronts him.”
    —Erich Fromm (1900–1980)