Spectral projected gradient method for conic optimization in kinematic limit analysis

0575725 - ÚGN 2024 RIV US eng C - Konferenční příspěvek (zahraniční konf.)
Pospíšil, L. - Sysala, Stanislav - Čermák, M.
Spectral projected gradient method for conic optimization in kinematic limit analysis.
Proceedings of the International Conference on Numerical Analysis and Applied Mathematics 2021 (ICNAAM-2021). Volume 2849, Issue 1. New York: AIP Publishing LLC., 2023, č. článku 310007. ISBN 978-0-7354-4589-5.
[International Conference on Numerical Analysis and Applied Mathematics 2021. Sheraton Rhodes Resort (GR), 20.09.2021-26.09.2021]
Grant CEP: GA ČR(CZ) GA19-11441S
Institucionální podpora: RVO:68145535
Klíčová slova: kinematic limit analysis problem * convex quadratic function * strip-footing benchmark problem
Obor OECD: Applied mathematics
https://pubs.aip.org/aip/acp/article/2849/1/310007/2909126

The paper is focused on solution of the kinematic limit analysis problem arising in geotechnical stability assessment. This problem may be classified as conic optimization and for its solution, the Augmented Lagrangian method with adaptive penalty is proposed to enforce equality constraints. The corresponding inner problem leads to the minimization of a convex quadratic function on a closed convex feasible set defined by separable cones and it is solved by projected gradients with spectral step-lengths. The efficiency of the suggested solver is demonstrated on a strip-footing benchmark problem.

Trvalý link: https://hdl.handle.net/11104/0346526