Resilient Consensus Control for Multi-Agent Systems: A Comparative Survey.

Area of Science:

• Computer Science
• Control Engineering
• Network Security

Background:

• Multi-agent systems (MAS) are susceptible to network attacks due to open communication and complex structures.
• These attacks can lead to significant instability in MAS operations.
• Understanding attack vectors is crucial for developing robust MAS.

Purpose of the Study:

• To provide a comprehensive survey of recent advances in network attacks targeting MAS.
• To review three primary attack types: Denial of Service (DoS), spoofing, and Byzantine attacks.
• To discuss attack mechanisms, models, and resilient consensus control structures.

Main Methods:

• Review of state-of-the-art research on network attacks in MAS.
• Detailed introduction of attack mechanisms for DoS, spoofing, and Byzantine attacks.
• Discussion of attack models and resilient consensus control structures, including theoretical innovations and limitations.

Main Results:

• Categorization and analysis of DoS, spoofing, and Byzantine attack mechanisms in MAS.
• Examination of the impact of these attacks on system stability and control.
• Presentation of existing results in a tutorial format for clarity.

Conclusions:

• Identified key challenges and open issues in resilient consensus control for MAS under network attacks.
• Highlighted the need for future research to address vulnerabilities and enhance system security.
• Provided guidance for future development directions in MAS security and resilience.