Вычислительные методы и приложения

An approach to solving inverse problems of computational diagnostics on supercomputers

A. V. Goncharsky

Lomonosov Moscow State University, Research Computing Center

Abstract: The paper is concerned with methods for solving inverse problems of mathematical physics on supercomputers. It discusses the approach based on the use of finite-dimensional parametric models in the cases of both deterministic and stochastic errors. Such methods are proved to be highly efficient when implemented on high-performance clusters.

Keywords: mathematical physics, inverse problems, parametric models, high-performance clusters, parallel computing.

UDC: 519.6