A stable fast time-stepping method for fractional integral and derivative operators
This work addresses the computational bottleneck of long-memory effects in fractional calculus for numerical simulations, offering a memory-efficient and accurate method.
The paper proposes a unified fast time-stepping method for fractional integral and derivative operators that decomposes the operator into local and history parts, achieving O(n_0 + sum q_α(N_ℓ)) active memory and O(n_0 n_T + (n_T - n_0) sum q_α(N_ℓ)) operations. The error from the truncated Laguerre-Gauss quadrature is shown to be independent of stepsize and arbitrarily small with suitable parameters.
A unified fast time-stepping method for both fractional integral and derivative operators is proposed. The fractional operator is decomposed into a local part with memory length $ΔT$ and a history part, where the local part is approximated by the direct convolution method and the history part is approximated by a fast memory-saving method. The fast method has $O(n_0+\sum_{\ell}^L{q}_α(N_{\ell}))$ active memory and $O(n_0n_T+ (n_T-n_0)\sum_{\ell}^L{q}_α(N_{\ell}))$ operations, where $L=\log(n_T-n_0)$, $n_0={ΔT}/τ,n_T=T/τ$, $τ$ is the stepsize, $T$ is the final time, and ${q}_α{(N_{\ell})}$ is the number of quadrature points used in the truncated Laguerre--Gauss (LG) quadrature. The error bound of the present fast method is analyzed. It is shown that the error from the truncated LG quadrature is independent of the stepsize, and can be made arbitrarily small by choosing suitable parameters that are given explicitly. Numerical examples are presented to verify the effectiveness of the current fast method.