Robust Adaptive Dynamic Programming for Large-Scale Systems With an Application to Multimachine Power Systems

This brief presents a new approach to decentralized control design of complex systems with unknown parameters and dynamic uncertainties. A key strategy is to use the theory of robust adaptive dynamic programming and the policy iteration technique. An iterative control algorithm is given to devise a decentralized optimal controller that globally asymptotically stabilizes the system in question. Stability analysis is accomplished by means of the small-gain theorem. The effectiveness of the proposed computational control algorithm is demonstrated via the online learning control of multimachine power systems with governor controllers.

I. Introduction

In recent years, considerable attention has been paid to the stabilization of large-scale complex systems [7], [17], [21]–[23], as well as the related consensus and synchronization problems [1], [16], [24], [34]. Examples of large-scale systems arise from ecosystems, transportation networks, and power systems, to name only a few [2], [15], [18], [20], [28], [36]. Often, in real-world applications, precise mathematical models are hard to build, and the model mismatches, caused by parametric and dynamic uncertainties, are thus unavoidable. This, together with the exchange of only local system information, makes the design problem extremely challenging in the context of complex networks.