Projection-based Prediction-Correction Method for Distributed Consensus Optimization
Han Long
Han Long
Sep 2023
0被引用
0笔记
开学季活动火爆进行中，iPad、蓝牙耳机、拍立得、键盘鼠标套装等你来拿
摘要原文
In the industrial technology domain, mathematical optimization is crucial with its applications seen in areas like transportation engineering, robotics, and machine learning. With the growth in data volume, there's an increased demand for solutions to large-scale challenges, leading to the rise of distributed optimization. This approach involves decentralized devices working collectively to achieve system objectives. The focus of the study is on distributed consensus optimization concerning convex set constraints in networks. The paper introduces the self-adaptive Projection-based Prediction-Correction Method (PPCM), inspired by the proximal method and integrated with variational inequality. PPCM stands out as a contractive method characterized by impressive convergence properties. Its decentralized nature also fits networked settings aptly. Also the parameter selection is simple and clear, without the hassle of parameter tuning. A thorough theoretical evaluation confirms PPCM's effectiveness. Upon applying the method to distributed linear least squares problems, it manifested a performance superiority, registering an enhancement exceeding 55% compared to Python's built-in functions. Overall, the research provides a robust distributed optimization technique with significant theoretical and practical benefits.
