Conference item
Safeguarded Anderson acceleration for parametric nonexpansive operators
- Abstract:
- This paper describes the design of a safeguarding scheme for Anderson acceleration to improve its practical performance and stability when used for first-order optimisation methods. We show how the combination of a nonexpansiveness condition, conditioning constraints, and memory restarts integrate well with solver algorithms that can be represented as fixed point operators with dynamically varying parameters. The performance of the scheme is demonstrated on seven different QP and SDP problem types, including more than 500 problems. The safeguarded Anderson acceleration scheme proposed in this paper is implemented in the opensource ADMM-based conic solver COSMO.
- Publication status:
- Published
- Peer review status:
- Peer reviewed
Actions
Access Document
- Files:
-
-
(Preview, Accepted manuscript, pdf, 370.2KB, Terms of use)
-
- Publisher copy:
- 10.23919/ECC55457.2022.9838320
Authors
- Publisher:
- IEEE
- Host title:
- Proceedings of the 20th European Control Conference (ECC 2022)
- Pages:
- 435-440
- Publication date:
- 2022-08-05
- Acceptance date:
- 2022-03-30
- Event title:
- 20th European Control Conference (ECC 2022)
- Event location:
- London, UK
- Event website:
- https://ecc22.euca-ecc.org/
- Event start date:
- 2022-07-12
- Event end date:
- 2022-07-15
- DOI:
- EISBN:
- 978-3-9071-4407-7
- ISBN:
- 978-1-6654-9733-6
- Language:
-
English
- Keywords:
- Pubs id:
-
1262906
- Local pid:
-
pubs:1262906
- Deposit date:
-
2022-06-10
Terms of use
- Copyright holder:
- IEEE
- Copyright date:
- 2022
- Rights statement:
- © IEEE 2022
- Notes:
- This paper will be presented at the 20th European Control Conference (ECC 2022), 12th-15th July 2022, London, UK. This is the accepted manuscript version of the article. The final version is available online from IEEE at: https://doi.org/10.23919/ECC55457.2022.9838320
If you are the owner of this record, you can report an update to it here: Report update to this record