Resilient Distributed Optimization With Event‐Triggered Interaction Design for Multiagent Systems Under False Data Injection Attacks

Abstract

ABSTRACT This article explores a novel design of a resilient interaction algorithm for multiagent systems (MAS) based on an event‐triggered mechanism, focusing on distributed optimization in the context of False Data Injection Attack (FDIA). A network‐level defense strategy is used based on a virtual system framework, where virtual state variables are introduced to ensure that the local estimate of each agent converges to the optimal solution of the distributed optimization problem, even under unknown FDIA. The article further introduces an event‐triggered strategy that significantly reduces communication overhead, and proper selection criteria are given for picking suitable event‐triggered parameters therein. It is proved that the proposed algorithm also avoids the Zeno behavior. Additionally, a distributed detection method is designed to accurately identify and isolate compromised links, thereby further enhancing the system's resilience. Two numerical simulations are conducted to illustrate the performance of the proposed algorithm, and it is demonstrated that the algorithm can also maintain effectiveness for networks with relatively large‐scale sizes.