Methods of incorporating a priori information into regularizing gradient algorithms

Methods are proposed for taking into account a priori information concerning the unknown quantity during the solution of boundary value and coefficient inverse heat conduction problems by gradient methods. As in the case of any ill-posed problems, the inclusion of a priori information, particularly information on the smoothness of the solution, makes it possible to significantly improve the quality of the resulting approximate solutions.