Abstract:
We consider the peculiarities of applying multiprocessor systems for solving inverse problems leading to multidimensional Fredholm integral equations of the first kind for a vector function. A parallel algorithm for solving problems of this kind is proposed for the two- and three-dimensional cases, which are the most general from the practical point of view. The efficiency of the proposed approach is shown by an example of solving the inverse problem of recovering magnetization parameters of a ship from measured values of the magnetic field outside its hull.
Keywords:inverse problems, ill-posed problems, Fredholm integral equation of the first kind, Tikhonov regularization method, parallel computing, recovery of magnetization parameters.
Fulltext:
PDF file (1503 kB)