The main aim of this paper is to construct an efficient highly accurate numerical scheme to solve a class of one and two-dimensional parabolic integro-fractional differential equations. The high order $L2$-$1_\sigma$ scheme is taken into account to discretize the time-fractional operator on a uniform mesh. To solve the one-dimensional problem, second-order discretizations are used to approximate the spatial derivatives whereas, a repeated quadrature rule based on trapezoidal approximation is employed to discretize the integral operator. In the case of two-dimensional problem, first, we make the semi-discretization of the proposed model based on the $L2$-$1_\sigma$ scheme for the fractional operator, and composite trapezoidal approximation for the integral part. Then, the spatial derivatives are approximated by the two-dimensional Haar wavelet. The stability and convergence analysis is carried out for both models. The experimental evidence proves the strong reliability of the present methods. Further, the obtained results are compared with some existing methods through several graphs and tables, and it is shown that the proposed methods not only have better accuracy but also produce less error in comparison with the $L1$ scheme.
翻译:本文的主要目标是构建一种高效精确的数值算法,用于解决一类一维和二维抛物型分数阶积分微分方程。 采用高阶的$L2$-$1_\sigma$方案来在均匀网格上离散时间分数阶算子。对于一维问题,采用二阶离散化来逼近空间导数,而采用基于梯形逼近的重复积分规则来离散化积分算子。对于二维问题,首先基于$L2$-$1_\sigma$方案对分数算子进行半离散化,对积分部分采用复合梯形逼近。然后,采用二维Haar小波逼近空间导数。对于两种模型,都进行了稳定性和收敛性分析。实验证据证明了目前方法的可靠性。此外,通过几个图表将所得结果与一些现有方法进行比较,表明本文方法不仅具有更好的精度,而且产生的误差比$L1$方案小。