Backend
fOGA: An Orthogonal Greedy Algorithm for Fractional Laplacian Problems
arXiv cs.NA0 views
arXiv:2409.16551v3 Announce Type: replace Abstract: We extend energy-based orthogonal greedy approximation to integral fractional Poisson problems by combining OGA with discrete fractional operators. In one dimension, order-dependent Gr\"unwald--Letnikov stencils are paired with boundary-corrected ReLU atoms. On the unit disk, a fractional centered-difference operator is constructed on a larger periodic lattice and restricted to the interior, while boundary-weighted ReLU$^2$ atoms impose the homogeneous exterior condition. The resulting symmetric positive definite matrices provide a common energy framework for normalized atom selection and orthogonal projection. We establish the corresponding discrete projection properties and separate operator discretization from approximation in the selected spaces. One- and two-dimensional experiments assess the method through $L^2$, $H^1$, and sampled maximum errors. The results demonstrate the applicability of the OGA extension to nonlocal problems and show the influence of boundary regularity on approximation, particularly in the $H^1$ error.
Source:https://arxiv.org/abs/2409.16551PreviousA Matching Lower Bound for Online Smooth CalibrationNextSynthetic Acceleration Preconditioners for Parametric Radiative Transfer Equations based on Trajectory-Aware Reduced-Order Models
Related posts
How zkDatabase Works: From Data Sampling to Verifiable Query Results
0dev.to Database
LogTreeDB: Building a Crash-Safe, Zero-Dependency Document Database from Scratch in C++17
0dev.to Database
CratesBump: A Version-Bump Dataset for the Rust Ecosystem
0arXiv cs.SE
Integrating AI Coding Agents Into Scientific Research
0arXiv cs.SE
Training for sustainable scientific software in the AI era: from code production to scientific stewardship
0arXiv cs.SE
A level is not a rank: an empirical audit of quantile-to-index conversion in 10 released conformal-prediction libraries
0arXiv cs.SE