PIAS in black-box optimization remains beneficial versus the single best algorithm for most tested cases even with 25% budget spent on features, and feature computation explains about 20% of the average performance gap to the virtual best solver.
Research Report RR-6829, INRIA (2009),https://inria.hal.science/inria-00362633 On the Influence of the Feature Computation Budget on PIAS for BBO 15
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.NE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
On the Influence of the Feature Computation Budget on Per-Instance Algorithm Selection for Black-Box Optimization
PIAS in black-box optimization remains beneficial versus the single best algorithm for most tested cases even with 25% budget spent on features, and feature computation explains about 20% of the average performance gap to the virtual best solver.