Scholarlink

<p>In this work, we investigate the distributed constrained online optimization problem based on continuous-time nonlinear multi-agent systems. The collective goal is to optimize the global objective at each time step while adhering to the coupling constraints. Every agent, however, can exclusively access its local cost function and constraint function at the present moment. To address this challenge, we introduce a comprehensive algorithmic framework that decomposes the problem into two steps: distributed online output optimization and trajectory tracking control. By applying an improved online optimization algorithm, the global cost function is minimized under coupling constraints and an estimated desired trajectory which is well-optimized is obtained at the same time. Subsequently, we employ the backstepping method to ensure that each agent’s state tracks the desired trajectory closely. We further demonstrate that a constant regret bound and a sublinear constraint violation bound are attained. To validate the theoretical findings, we conducted simulations with a one-link manipulator system, showcasing the efficacy of the proposed approach.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Distributed continuous-time constrained online convex optimization for high-order nonlinear multi-agent systems

  • 期刊论文

  • 发表日期：

    2026年5月18日

查看全文
  • Yao Yao,
  • Yuxiao Lian,
  • Baoyong Zhang,
  • Deming Yuan,
  • Bo Song

摘要

In this work, we investigate the distributed constrained online optimization problem based on continuous-time nonlinear multi-agent systems. The collective goal is to optimize the global objective at each time step while adhering to the coupling constraints. Every agent, however, can exclusively access its local cost function and constraint function at the present moment. To address this challenge, we introduce a comprehensive algorithmic framework that decomposes the problem into two steps: distributed online output optimization and trajectory tracking control. By applying an improved online optimization algorithm, the global cost function is minimized under coupling constraints and an estimated desired trajectory which is well-optimized is obtained at the same time. Subsequently, we employ the backstepping method to ensure that each agent’s state tracks the desired trajectory closely. We further demonstrate that a constant regret bound and a sublinear constraint violation bound are attained. To validate the theoretical findings, we conducted simulations with a one-link manipulator system, showcasing the efficacy of the proposed approach.