Back to Search
Start Over
Finding Near-optimal Configurations in Colossal Spaces with Statistical Guarantees.
- Source :
- ACM Transactions on Software Engineering & Methodology; Jan2024, Vol. 33 Issue 1, p1-36, 36p
- Publication Year :
- 2024
-
Abstract
- A Software Product Line (SPL) is a family of similar programs. Each program is defined by a unique set of features, called a configuration, that satisfies all feature constraints. "What configuration achieves the best performance for a given workload?" is the SPLOptimization (SPLO) challenge. SPLO is daunting: just 80 unconstrained features yield 10<superscript>24</superscript> unique configurations, which equals the estimated number of stars in the universe. We explain (a) how uniform random sampling and random search algorithms solve SPLO more efficiently and accurately than current machine-learned performance models and (b) how to compute statistical guarantees on the quality of a returned configuration; i.e., it is within x% of optimal with y% confidence. [ABSTRACT FROM AUTHOR]
- Subjects :
- PRODUCT lines
STATISTICAL sampling
STRUCTURAL optimization
ORDER statistics
Subjects
Details
- Language :
- English
- ISSN :
- 1049331X
- Volume :
- 33
- Issue :
- 1
- Database :
- Complementary Index
- Journal :
- ACM Transactions on Software Engineering & Methodology
- Publication Type :
- Academic Journal
- Accession number :
- 174912631
- Full Text :
- https://doi.org/10.1145/3611663