ParBalans: Parallel Multi-Armed Bandits-based Adaptive Large Neighborhood Search
Fuente:
arXiv
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911100333719552 |
|---|---|
| author | Yilmaz, Alican Cai, Junyang Kadioglu, Serdar Dilkina, Bistra |
| author_facet | Yilmaz, Alican Cai, Junyang Kadioglu, Serdar Dilkina, Bistra |
| contents | Solving Mixed-Integer Programming (MIP) problems often requires substantial computational resources due to their combinatorial nature. Parallelization has emerged as a critical strategy to accelerate solution times and enhance scalability to tackle large, complex instances. This paper investigates the parallelization capabilities of Balans, a recently proposed multi-armed bandits-based adaptive large neighborhood search for MIPs. While Balans's modular architecture inherently supports parallel exploration of diverse parameter configurations, this potential has not been thoroughly examined. To address this gap, we introduce ParBalans, an extension that leverages both solver-level and algorithmic-level parallelism to improve performance on challenging MIP instances. Our experimental results demonstrate that ParBalans exhibits competitive performance compared to the state-of-the-art commercial solver Gurobi, particularly on hard optimization benchmarks. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_06736 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | ParBalans: Parallel Multi-Armed Bandits-based Adaptive Large Neighborhood Search Yilmaz, Alican Cai, Junyang Kadioglu, Serdar Dilkina, Bistra Artificial Intelligence Machine Learning Solving Mixed-Integer Programming (MIP) problems often requires substantial computational resources due to their combinatorial nature. Parallelization has emerged as a critical strategy to accelerate solution times and enhance scalability to tackle large, complex instances. This paper investigates the parallelization capabilities of Balans, a recently proposed multi-armed bandits-based adaptive large neighborhood search for MIPs. While Balans's modular architecture inherently supports parallel exploration of diverse parameter configurations, this potential has not been thoroughly examined. To address this gap, we introduce ParBalans, an extension that leverages both solver-level and algorithmic-level parallelism to improve performance on challenging MIP instances. Our experimental results demonstrate that ParBalans exhibits competitive performance compared to the state-of-the-art commercial solver Gurobi, particularly on hard optimization benchmarks. |
| title | ParBalans: Parallel Multi-Armed Bandits-based Adaptive Large Neighborhood Search |
| topic | Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2508.06736 |