Saved in:
| Main Authors: | , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2506.04509 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912414646140928 |
|---|---|
| author | Ganguly, Kishan Kumar Menzies, Tim |
| author_facet | Ganguly, Kishan Kumar Menzies, Tim |
| contents | Traditional multi-objective optimization in software engineering (SE) can be slow and complex. This paper introduces the BINGO effect: a novel phenomenon where SE data surprisingly collapses into a tiny fraction of possible solution "buckets" (e.g., only 100 used from 4,096 expected).
We show the BINGO effect's prevalence across 39 optimization in SE problems. Exploiting this, we optimize 10,000 times faster than state-of-the-art methods, with comparable effectiveness. Our new algorithms (LITE and LINE), demonstrate that simple stochastic selection can match complex optimizers like DEHB. This work explains why simple methods succeed in SE-real data occupies a small corner of possibilities-and guides when to apply them, challenging the need for CPU-heavy optimization.
Our data and code are public at GitHub (see anon-artifacts/bingo). |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_04509 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | BINGO! Simple Optimizers Win Big if Problems Collapse to a Few Buckets Ganguly, Kishan Kumar Menzies, Tim Software Engineering Traditional multi-objective optimization in software engineering (SE) can be slow and complex. This paper introduces the BINGO effect: a novel phenomenon where SE data surprisingly collapses into a tiny fraction of possible solution "buckets" (e.g., only 100 used from 4,096 expected). We show the BINGO effect's prevalence across 39 optimization in SE problems. Exploiting this, we optimize 10,000 times faster than state-of-the-art methods, with comparable effectiveness. Our new algorithms (LITE and LINE), demonstrate that simple stochastic selection can match complex optimizers like DEHB. This work explains why simple methods succeed in SE-real data occupies a small corner of possibilities-and guides when to apply them, challenging the need for CPU-heavy optimization. Our data and code are public at GitHub (see anon-artifacts/bingo). |
| title | BINGO! Simple Optimizers Win Big if Problems Collapse to a Few Buckets |
| topic | Software Engineering |
| url | https://arxiv.org/abs/2506.04509 |