Skip to main navigation Skip to search Skip to main content

Temporal second-order H1-norm convergence of the integral-averaged L1 method for time-fractional mobile–immobile diffusion equations

  • Shanghai Normal University Tianhua College

Research output: Contribution to journalArticlepeer-review

Abstract

This paper studies the integral-averaged L1 (IAL1) method of the Caputo time-fractional derivative of order α∈(0,1). Under a certain singular regularity condition, Zheng and Wang (2024) proved that on uniform time meshes, the IAL1 method for time-fractional mobile–immobile (MIM) diffusion equations has a temporal convergence order of min{3−2α,2}, which means that when α>1/2, the IAL1 method cannot achieve temporal second-order convergence on uniform time meshes. Furthermore, the numerical experiments by Zheng and Wang (2024) indicate that on a suitably graded time mesh, the IAL1 method can achieve temporal second-order convergence even when α>1/2, but the corresponding convergence analysis remains unavailable. To fill this gap, under relatively weak regularity conditions, we give an H1-norm error estimate for the IAL1 method of time-fractional MIM diffusion equations on graded time meshes. For certain cases of initially weakly singular solutions, we prove that the method can achieve temporal second-order convergence under a weaker mesh grading constraint for all α∈(0,1). Our proof relies on the positive definiteness of the IAL1 derivative operator established on general time meshes. Numerical results confirm the theoretical results.

Original languageEnglish
Article number110021
JournalApplied Mathematics Letters
Volume181
DOIs
StatePublished - Oct 2026

Keywords

  • Graded time meshes
  • H-norm convergence
  • Integral-averaged L1 method
  • Positive definiteness
  • Time-fractional mobile–immobile diffusion equation

Fingerprint

Dive into the research topics of 'Temporal second-order H1-norm convergence of the integral-averaged L1 method for time-fractional mobile–immobile diffusion equations'. Together they form a unique fingerprint.

Cite this