JournalVolume 12, №2Gradient and Direct Method of Solving Gelfand-Levitan Integral Equation (pages 234-246) L.N. Temirbekova, G. Dairbaeva Abstract. In this work, we consider numerical methods of solving inverse problems for hyperbolic equations. It is necessary to restore the coefficient of extra information for solving direct problems in the reverse one. One of the initial problem specifications of the direct problem is the Dirac delta function. Taking into consideration the peculiarity, the direct problem is replaced by the Goursat problem, with classical derivatives. The Goursat problem with the data on characteristics, is numerically solved using the finite-difference method. The coefficient inverse problem is numerically solved using the method of the Gelfand-Levitan integral equation. To numerically solve the Gelfand-Levitan integral equation, which is the Fredholm integral equation of first kind, they use the Lavrentiev regularization method, the direct square root method, and the conjugate gradient method. We present the effectiveness of the developed numerical conjugate gradient method.
Keywords: Equation, Difference Scheme, Approximation, Stability and Convergence, Regularization Method, Gradient Methods, Error Norm.
AMS Subject Classification: 65M32.
|
|