Distributed fault-tolerant consensus control for multi-agent systems under complex fault dynamics using game theory and sparse fuzzy hyperbolic models
Qiuyu Miao, Zhigang Wu · International Journal of Systems Science · 2025
This paper presents a distributed fault-tolerant optimal consensus control framework for large-scale multi-agent systems (MAS) under complex fault dynamics. The proposed approach employs a novel hybrid policy-value iteration algorithm based on a generalised fuzzy hyperbolic model (GFHM) for Nash equilibrium solutions in nonzero-sum games. To enhance computational efficiency, the framework incorporates Sparse Adaptive Optimization with the GFHM structure to promote weight sparsity while maintaining control performance. The framework handles complex hybrid fault dynamics through a semi-Markov model with time-dependent transitions, advancing traditional Markov approaches that assume constant failure rates. A fully distributed consensus protocol with enhanced scalability is established through dynamic agent addition/removal capabilities, eliminating the need for global reconfiguration. Experimental validation on MAS demonstrates the framework's effectiveness in maintaining system stability and performance under varying fault conditions, while achieving improved computational efficiency and reduced communication overhead compared to traditional approaches.