Distributed Optimization for Heterogeneous MASs Under Unbalanced Topologies and DoS Attacks

Mali Xing

doi:10.1109/tase.2026.3708986

Distributed Optimization for Heterogeneous MASs Under Unbalanced Topologies and DoS Attacks

2026-07-01

0

PRE

AI

DOI：10.1109/tase.2026.3708986

Original

Original request for help

Save

Abstract

En 中文

This paper investigates the distributed optimization problem (DOP) for heterogeneous linear multi-agent systems (MASs) with weight-unbalanced directed communication networks subject to denial-of-service (DoS) attacks. These malicious DoS attacks may disrupt some communication links preventing the transmission of measurement information. A distributed optimization controller algorithm, integrating event-triggered mechanisms (ETMs) and a compensator, is proposed to alleviate communication congestion and solve the problem caused by the umbalance of the communication topology and malicious DoS attacks. It shows that the well designed optimization controller can tolerate a type of DoS signals that characterized by the duration and frequency of the attacks. The proposed control strategy significantly reduces communication frequency while ensuring global exponential stability and convergence to the optimal solution. The theoretical analysis excludes Zeno behavior, and numerical simulations validate the effectiveness and robustness of the approach under intermittent DoS attacks. Note to Practitioners—In many engineering systems—such as smart grids, UAV formations, and distributed robotics—agents must optimize global objectives despite limited communication and exposure to cyber threats. In practice, these systems often operate with different local dynamics, unbalanced information flow, and intermittent denial-of-service (DoS) attacks that block data transmission. Existing algorithms seldom address these constraints simultaneously. This work provides a practical and implementable solution for such environments. The proposed controller compensates for network imbalance, tolerates DoS attacks within measurable bounds, and uses event-triggered updates to reduce communication load. The method ensures stable convergence without requiring continuous communication or homogeneous agent models. Practitioners can directly apply these design principles to improve the resilience and efficiency of real-world distributed control systems.

Keywords:

Distributed optimization

denial-of-service attacks

event-triggered mechanism

multi-agent system (MAS)

Distributed Optimization for Heterogeneous MASs Under Unbalanced Topologies and DoS Attacks

Abstract

Journal

Organization