arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2303.11264eess.SYcs.SYmath.OC
收藏

局部模型预测控制的全局性能保证

Global Performance Guarantees for Localized Model Predictive Control

Jing Shuang Li, Carmen Amo Alonso

AI总结:

本文为局部模型预测控制提供全局性能保证,推导最优性能的充分条件,并提出最小化计算复杂度的通信结构算法,证明121个耦合摆网络中仅需与直接邻居通信即可保持最优性能。

摘要:

模型预测控制(MPC)的最新进展利用局部通信约束来产生局部化MPC算法,其复杂度与网络总规模无关。然而,关于全局性能的表征尚不可用,即局部化MPC(带通信约束)是否与全局MPC(无通信约束)表现同样出色。在本文中,我们提供了对局部化MPC全局性能的分析和保证——特别是,我们推导了在存在局部通信约束时实现最优全局性能的充分条件。我们还提出了一种算法,用于确定给定系统的通信结构,该结构将在保持性能的同时最小化计算复杂度。该算法的有效性在仿真中得到了验证,并表征了网络属性与保持性能的通信约束之间的额外关系。一个显著的发现是,在121个耦合摆的网络中,每个节点仅需与其直接邻居通信即可保持最优全局性能。总体而言,这项工作提供了关于局部通信对全局性能影响的理论理解,并通过建立选择局部通信约束的严谨方法,为实践者提供了部署局部化模型预测控制所需的工具。这项工作还表明——令人惊讶的是——包含严格的通信约束并不一定损害全局性能。

英文摘要:

Recent advances in model predictive control (MPC) leverage local communication constraints to produce localized MPC algorithms whose complexities scale independently of total network size. However, no characterization is available regarding global performance, i.e. whether localized MPC (with communication constraints) performs just as well as global MPC (no communication constraints). In this paper, we provide analysis and guarantees on global performance of localized MPC -- in particular, we derive sufficient conditions for optimal global performance in the presence of local communication constraints. We also present an algorithm to determine the communication structure for a given system that will preserve performance while minimizing computational complexity. The effectiveness of the algorithm is verified in simulations, and additional relationships between network properties and performance-preserving communication constraints are characterized. A striking finding is that in a network of 121 coupled pendula, each node only needs to communicate with its immediate neighbors to preserve optimal global performance. Overall, this work offers theoretical understanding on the effect of local communication on global performance, and provides practitioners with the tools necessary to deploy localized model predictive control by establishing a rigorous method of selecting local communication constraints. This work also demonstrates -- surprisingly -- that the inclusion of severe communication constraints need not compromise global performance.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑