Distributed Protocols for Constrained Optimization of Integrator Chain Multiagent Systems

• 成果类型: Article
• 署名作者: Zou, Yao; Huang, Bomin; Sun, Yongbin; Li, Qing; Meng, Ziyang; He, Wei
• 署名单位: University of Science & Technology Beijing; University of Science & Technology Beijing; University of Science & Technology Beijing; Jimei University; Beihang University; Tsinghua University
• 刊物名称: IEEE TRANSACTIONS ON AUTOMATIC CONTROL
• ISSN/ISSBN: 0018-9286
• DOI: 10.1109/TAC.2024.3394128
• 发表日期: 2024
• 页码: 7197-7204
• 关键词: Optimization; Protocols; Multi-agent systems; linear programming; stem; Aggregates; Decision feedback equalizers; constrained optimization; distributed protocol; equality constraint; integrator chain system; set constraint

摘要

This article studies the constrained optimization problem of integrator chain multiagent systems from the distributed perspective. Particularly, all the agents collaborate to find an optimal solution of a global objective function summed by multiple local ones, each of which is just convex and belongs to an individual agent. However, the potential optimal solutions are simultaneously subject to set constraints and two kinds of equality constraints, i.e., individual equality constraint and aggregate equality constraint, whereas their involved constraint parameters are privately available to individual agents. With a proper coordinate transformation, a distributed optimization protocol is synthesized by introducing proper internal dynamics. Under the Lyapunov framework, it is shown with a convex theory that the synthesized distributed protocol enables the multiagent system to achieve the concerned constrained optimization objective. An example is finally given to confirm the optimization effectiveness.