Cited by
- BibTex
- RIS
- TXT
The numerical solution of the fractional sub-diffusion equations on one dimensional space unbounded domain is considered. Based on the high-order local artificial boundary conditions proposed in [Zhang W., et al., J. Math. Study., 2017, 50(1): 28-53], the original space unbounded problem can be reformulated to an initial-boundary value problem on a bounded computational domain. By Alikhanov’s $L2-1_σ$ formula and sum-of-exponentials approximation, a fast temporal second order difference scheme for the reduced problem is presented. The unique solvability, stability and convergence order $O(τ^2+h^2 )$ of the proposed method are proved by means of energy method, where $τ$ and $h$ denote the time and space step sizes, respectively. Some numerical examples are included to validate the theoretical results. To the best of our knowledge, this is the first work that combines the high order numerical method with the artificial boundary method for the time fractional diffusion problems on spatial unbounded domains.
}, issn = {2617-8702}, doi = {https://doi.org/10.4208/jms.v56n2.23.05}, url = {http://global-sci.org/intro/article_detail/jms/21837.html} }The numerical solution of the fractional sub-diffusion equations on one dimensional space unbounded domain is considered. Based on the high-order local artificial boundary conditions proposed in [Zhang W., et al., J. Math. Study., 2017, 50(1): 28-53], the original space unbounded problem can be reformulated to an initial-boundary value problem on a bounded computational domain. By Alikhanov’s $L2-1_σ$ formula and sum-of-exponentials approximation, a fast temporal second order difference scheme for the reduced problem is presented. The unique solvability, stability and convergence order $O(τ^2+h^2 )$ of the proposed method are proved by means of energy method, where $τ$ and $h$ denote the time and space step sizes, respectively. Some numerical examples are included to validate the theoretical results. To the best of our knowledge, this is the first work that combines the high order numerical method with the artificial boundary method for the time fractional diffusion problems on spatial unbounded domains.