Mixed-Zero-Sum-Game-Based Memory Event-Triggered Cooperative Control of Heterogeneous MASs Against DoS Attacks.

Area of Science:

• Control Systems Engineering
• Artificial Intelligence
• Network Security

Background:

• Heterogeneous nonlinear multiagent systems (MASs) are susceptible to denial-of-service (DoS) attacks, disrupting cooperative control.
• Traditional control strategies struggle with dynamic network topologies and external interference.
• Event-triggered mechanisms are needed to optimize communication and response under uncertain conditions.

Purpose of the Study:

• To develop a memory event-triggered cooperative adaptive control strategy for heterogeneous nonlinear MASs under DoS attacks.
• To address the multiplayer mixed zero-sum (ZS) game problem in MASs with switching topologies and node-based DoS attacks.
• To ensure system stability and leader-following performance despite intermittent attacks.

Main Methods:

• A neural-network-based reinforcement learning scheme is employed to find the Nash equilibrium of the multiplayer mixed ZS game.
• A novel memory-based event-triggered mechanism utilizing historical data is proposed to prevent false triggers.
• Lyapunov stability theory is applied to analyze system boundedness and control performance.

Main Results:

• All signals in the heterogeneous MASs are proven to be bounded under the proposed control strategy.
• Followers achieve trajectory tracking during no-attack periods and stabilization/cooperative control during attack periods.
• The memory-based event-triggered mechanism effectively mitigates the impact of DoS attacks and external factors.

Conclusions:

• The proposed memory-event-triggered cooperative adaptive control strategy based on a mixed ZS game is effective for heterogeneous MASs under DoS attacks.
• The integration of neural networks, event-triggering, and switching topologies provides robust control solutions.
• Simulation results validate the strategy's ability to maintain stability and achieve cooperative control objectives.