Data-Driven Model-Free Optimal Cooperative Control of Multi-Agent Systems

This paper introduces an innovative methodology based on data-driven for addressing collaborative control in multi-agent systems. A nonparametric representation model of the considered systems is first constructed by leveraging offline data. Subsequently, a distributed consensus controller is designed using data-driven methods, which dynamically adjusts control strategies through online interaction data without requiring explicit system models. For optimal control requirements in multi-agent systems, a data-driven linear quadratic regulator solution is developed. This approach determines the optimal control law by solving the Riccati equation to obtain the state feedback gain matrix, ensuring system stability while minimizing the performance index. The efficacy of the developed approach is further validated through simulation experiments.