2025-11-03 マサチューセッツ工科大学（MIT）
Web要約 の発言：
＜関連情報＞
- https://news.mit.edu/2025/faster-problem-solving-tool-guarantees-feasibility-1103
- https://arxiv.org/abs/2506.00362
- https://arxiv.org/pdf/2506.00362
FSNet: 保証付き制約最適化のための実現可能性探索ニューラルネットワーク FSNet: Feasibility-Seeking Neural Network for Constrained Optimization with Guarantees
Hoang T. Nguyen, Priya L. Donti
arXiv ast revised 24 Oct 2025 (this version, v2)
DOI:https://doi.org/10.48550/arXiv.2506.00362
Abstract
Efficiently solving constrained optimization problems is crucial for numerous real-world applications, yet traditional solvers are often computationally prohibitive for real-time use. Machine learning-based approaches have emerged as a promising alternative to provide approximate solutions at faster speeds, but they struggle to strictly enforce constraints, leading to infeasible solutions in practice. To address this, we propose the Feasibility-Seeking Neural Network (FSNet), which integrates a feasibility-seeking step directly into its solution procedure to ensure constraint satisfaction. This feasibility-seeking step solves an unconstrained optimization problem that minimizes constraint violations in a differentiable manner, enabling end-to-end training and providing guarantees on feasibility and convergence. Our experiments across a range of different optimization problems, including both smooth/nonsmooth and convex/nonconvex problems, demonstrate that FSNet can provide feasible solutions with solution quality comparable to (or in some cases better than) traditional solvers, at significantly faster speeds.
