Abstract
|
In this paper, we study the LS-algorithm for solving linear integral equations of
the first kind. This method is based on the reducing the solution of first kind linear
integral equations to the solution of a least squares problem with bidiagonal matrix.
Then applying the QR factorization method leads to a simple recurrence formula for
generating the sequence of approximate solutions. Some properties and convergence
theorem are proposed. Moreover, regularization property of the new method with a
suitable stopping rule is studied. Finally, some numerical examples are presented to
show the efficiency of the new method.
|