Distributed Optimization of Linear Systems Based on Data-Driven
Zhiwei Zhang, Yi Zhang · 2025
In this paper, the distributed optimization problem of a class of continuous time multi-agent systems with linear dynamics with unknown parameters is studied. Considering a given global convex objective function, the objective of this paper is that each agent guides the state of the multi-agent system to the optimal solution of the global objective function using only the local interaction and the gradient of its own local objective function. In order to achieve this goal, the input data and state response are used to reconstruct the system to estimate the parameter matrix. A data-driven distributed optimization algorithm is proposed based on the reconstructed system to ensure that the state of all agents can follow the global optimal solution. Then the stability of the whole closed-loop system composed of data-driven multi-agent and data-driven distributed protocol is established. Sufficient conditions are given to ensure that all agents achieve state consistency while minimizing the objective function. Finally, a numerical example is given to verify the validity of the theoretical results.