Текущий выпуск Выпуск 1, 2025 Том 35
Результыты поиска по 'grid schemes':
Найдено статей: 2
  1. Рассматривается уравнение в частных производных первого порядка с эффектом наследственности:

    $$ \frac{\partial u(x,t)}{\partial t} + a \frac{\partial u(x,t)}{\partial x} = f ( x, t, u(x,t), u_t(x,\cdot)),$$ $$u_t(x,\cdot) = \{u(x,t+s), -\tau\leqslant s <0\}.$$

    Для такого уравнения, с позиций принципа разделения конечномерной и бесконечномерной составляющих состояния, строятся сеточные методы: аналог семейства схем бегущего счета, аналог схемы Кранка-Николсон, метод аппроксимации на середину квадрата. Для учета эффекта наследственности применяются одномерная и двойная кусочно-линейная интерполяции и экстраполяция продолжением. Доказывается, что рассмотренные методы имеют порядки локальной погрешности: соответственно $O(h+\Delta)$, $O(h+\Delta^2)$ и $O(h^2+\Delta^2)$, где $h$ - шаг дискретизации по пространственной переменной, $\Delta$ - шаг дискретизации по временной переменной. Исследуются свойства двойной кусочно-линейной интерполяции. Используя результаты общей теории разностных схем, установлены условия устойчивости предложенных методов. С помощью вложения в общую схему численных методов для функционально-дифференциальных уравнений получены теоремы о порядках сходимости сконструированных алгоритмов. Приведены тестовые примеры по сравнению погрешностей методов.

    Pimenov V.G., Sviridov S.V.
    Grid methods of solving advection equations with delay, pp. 59-74

    We consider a first-order partial differential equation with heredity effect

    $$ \frac{\partial u(x,t)}{\partial t} + a \frac{\partial u(x,t)}{\partial x} = f ( x, t, u(x,t), u_t(x,\cdot)),$$ $$u_t(x,\cdot) = \{u(x,t+s), -\tau\leqslant s <0\}.$$

    For such an equation we construct grid methods using the principle of separation of finite-dimensional and infinite-dimensional state components. These grid methods are: analog of running schemes family, analog of Crank-Nicolson scheme, an approximation method to the middle of the square. The one-dimensional and double piecewise linear interpolation and the extrapolation by continuation are applied in order to account the effect of heredity. It is shown that the considered methods have orders of a local error: $O (h +\Delta) $, $O (h +\Delta^2) $ and $O (h^2 +\Delta^2)$ respectively, where $h$ is the spatial discretization interval, $\Delta$ is the time discretization interval. Properties of double piecewise linear interpolation are investigated. Using the results of the general theory of differential schemes, stability conditions of the proposed methods are established. Including them in the general scheme of numerical methods for the functional-differential equations, theorems of orders of proposed algorithms convergence are received. Test examples comparing errors of methods are given.

  2. Рассмотрено волновое уравнение с двумя пространственными и одной временной независимыми переменными и эффектом наследственности вида $$\frac{\partial^2 u}{\partial t^2}=a^2\left(\frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2}\right) + f\big(x,y,t,u(x,y,t),u_t(x,y,\cdot)\big),\\u_t(x,y,\cdot)=\left\{u(x,y,t+\xi),-\tau \leqslant \xi\leqslant 0\right\}. $$На основе идеи разделения текущего состояния и функции-предыстории сконструировано семейство сеточных методов для численного решения этого уравнения. По текущему состоянию строится полный аналог известного для уравнения без запаздывания метода с факторизацией, а влияние предыстории учитывается с помощью интерполяционных конструкций. Исследован порядок локальной погрешности алгоритма. Получена теорема о сходимости и порядке сходимости методов с помощью вложения в общую разностную схему систем с последействием. Приводятся результаты расчетов тестового примера с переменным запаздыванием.

    The paper presents the consideration of the wave equation with two space variables and one time variable and with heredity effect  $$\frac{\partial^2 u}{\partial t^2}=a^2\left(\frac{\partial^2 u}{\partial x^2} + \frac{\partial^2 u}{\partial y^2}\right) + f\big(x,y,t,u(x,y,t),u_t(x,y,\cdot)\big),\\u_t(x,y,\cdot)=\left\{u(x,y,t+\xi),-\tau \leqslant \xi\leqslant 0\right\}. $$A family of grid methods is constructed for the numerical solution of this equation; the methods are based on the idea of separating the current state and the history function. A complete analog of the factorization method which is known for an equation without delay is constructed according to the current state. Influence of prehistory is taken into consideration by interpolation constructions. The local error order of the algorithm is investigated. A theorem on the convergence and on the order of convergence of methods is obtained by means of embedding into a general difference scheme with aftereffect. The results of calculating a test example with variable delay are presented.

Журнал индексируется в Web of Science (Emerging Sources Citation Index)

Журнал индексируется в Scopus

Журнал входит в базы данных zbMATH, MathSciNet

Журнал включен в базу данных Russian Science Citation Index (RSCI) на платформе Web of Science

Журнал включен в перечень ВАК.

Электронная версия журнала на Общероссийском математическом портале Math-Net.Ru.

Журнал включен в Crossref