Decentralized output feedback guaranteed cost control of uncertain markovian jump large-scale systems: Local mode dependent control approach

Large-scale systems provide a mathematical model to represent dynamic systems that have a complex structure and high dimensions. Many physical systems, such as power systems [16], can be modeled as large-scale systems. In a large-scale system setting, one often lacks the system information that is necessary to implement a centralized controller design, or the complexity of a centralized design and cost of implementing centralized controllers is prohibitive. In such cases, decentralized controllers that only use locally available subsystem information provide a viable alternative to centralized solutions. In many practical problems involving large-scale systems, the decentralized control technique can effectively deal with such issues as information structure constraints, large dimensions and a broad range of uncertainties. However, the design of decentralized controllers is challenging for at least two reasons. Firstly, not all the system information is available to the controllers. Secondly, the dynamics of each subsystem in the large-scale system are affected by other subsystems [19] via the interconnections; in many situations the effect of those interconnections is essentially uncertain.