Open Distributed Convex Optimization

• 成果类型: Article
• 署名作者: Wu, Xiaoyu; Lin, Jie; Jiang, Shanchan; Zhang, Zhicheng; Du, Yingxue; Zhang, Yan
• 署名单位: Zhejiang Normal University; Tianjin University
• 刊物名称: IEEE TRANSACTIONS ON AUTOMATIC CONTROL
• ISSN/ISSBN: 0018-9286
• DOI: 10.1109/TAC.2025.3608057
• 发表日期: 2026
• 关键词: continuous-time; resource-allocation; algorithm; consensus

摘要

This article investigates the distributed convex optimization problem where dimension of the feasible set is fluctuated. To address this, we formulate an open consensus algorithm that enables agents to join or leave at any time while seeking the optimal solution. By introducing an auxiliary variable to account for agent departures and incorporating a diminishing gradient descent scheme, we seek the Lyapunov-based method to establish conditions ensuring that consensus remains achievable and rigorously proving that agents aggregate to the optimal solution. Finally, we validate the proposed method and theoretical findings through both a numerical example and a practical instance.