Parallel sweep algorithm for solving direct and inverse problems for time-fractional diffusion equation

2022

Full text Open access: Hybrid Metadata (CSL-JSON)
DOI 10.26089/nummet.v23r417
Data confidence 70 medium
How this score is calculated

It is built from checkable properties of the record itself: whether the article has a DOI, how many independent sources confirmed its data, whether the required fields are filled in, and how recent the latest confirmation is. Disagreement between sources and a missing identifier lower it. A value of 80 or above means the record was published without further checking; 60 to 79 means it is published at reduced confidence, with a manual check advisable. A record verified by an editor by hand is never pushed below 90 by a recalculation. This scores our data about the article, not the article itself.

Abstract

Работа посвящена построению параллельных алгоритмов решения прямой начально-краевой задачи и обратной задачи о восстановлении правой части для уравнения диффузии с дробной производной по времени. При использовании дополнительной информации о решении в некоторой внутренней точке обратная задача сводится к прямой задаче для вспомогательного уравнения. После применения конечно-разностных схем задачи сводятся к решению систем линейных алгебраических уравнений. Разработанные алгоритмы основаны на методе параллельной прогонки и реализованы для многоядерных процессоров с использованием технологии OpenMP. Проведены численные эксперименты для исследования производительности разработанных алгоритмов. The work is devoted to construction of parallel algorithm for solving the direct initial boundary and inverse right-hand part identification problems for the time-fractional diffusion equation. Application of a priori information on the solution at the some inner point allows one to reduce the inverse problem to an initial boundary problem for the auxiliary equation. After applying the finite-difference scheme the problems are reduced to solving systems of linear algebraic equations. The developed algorithms are based on the parallel sweep method and implemented for the multicore processor using the OpenMP technology. Numerical experiments were performed to study the performance of the developed algorithms.