Open Collections will undergo maintenance on Thursday, July 24th, 2025. The site will not be available from 8:00 AM - 9:00 AM PST and performance may be impacted from 9:00 AM - 12:00 PM PST.
- Library Home /
- Search Collections /
- Open Collections /
- Browse Collections /
- BIRS Workshop Lecture Videos /
- The flow limit of reflect-reflect-relax
Open Collections
BIRS Workshop Lecture Videos
BIRS Workshop Lecture Videos
The flow limit of reflect-reflect-relax Elser, Veit
Description
In the reflect-reflect-average (RRA) iteration scheme a point is averaged with the product of its reflection in two constraint sets to produce the next point. It was noticed in a recent study of the bit retrieval problem that the relaxed variant of RRA, called RRR, has superior performance in the limit of zero relaxation parameter. In this limit RRR defines a continuous flow (a system of ODEs), where the flow field is formed by the difference of a pair of points on the constraint sets. The flow limit can be implemented very efficiently (event-chain dynamics) when the flow field is piecewise constant, as it is in some NP-hard feasibility problems. I hope to have results for the Latin square completion problem in time for the workshop.
Item Metadata
| Title |
The flow limit of reflect-reflect-relax
|
| Creator | |
| Publisher |
Banff International Research Station for Mathematical Innovation and Discovery
|
| Date Issued |
2017-09-21T09:00
|
| Description |
In the reflect-reflect-average (RRA) iteration scheme a point is averaged with the product of its reflection in two constraint sets to produce the next point. It was noticed in a recent study of the bit retrieval problem that the relaxed variant of RRA, called RRR, has superior performance in the limit of zero relaxation parameter. In this limit RRR defines a continuous flow (a system of ODEs), where the flow field is formed by the difference of a pair of points on the constraint sets. The flow limit can be implemented very efficiently (event-chain dynamics) when the flow field is piecewise constant, as it is in some NP-hard feasibility problems. I hope to have results for the Latin square completion problem in time for the workshop.
|
| Extent |
37 minutes
|
| Subject | |
| Type | |
| File Format |
video/mp4
|
| Language |
eng
|
| Notes |
Author affiliation: Cornell University
|
| Series | |
| Date Available |
2018-03-26
|
| Provider |
Vancouver : University of British Columbia Library
|
| Rights |
Attribution-NonCommercial-NoDerivatives 4.0 International
|
| DOI |
10.14288/1.0364463
|
| URI | |
| Affiliation | |
| Peer Review Status |
Unreviewed
|
| Scholarly Level |
Faculty
|
| Rights URI | |
| Aggregated Source Repository |
DSpace
|
Item Media
Item Citations and Data
Rights
Attribution-NonCommercial-NoDerivatives 4.0 International