Robust Model Predictive Control Of Markov Jump System

Stochastic systems are used to describe the phenomenon of random systems: such as wind power when the wind strength,the flight encountered encountered changes in air flow,signal propagation disturbance and so on.For these random phenomena,we can establish the corresponding stochastic differential equations,which can be described by stochastic systems.In recent years,the problem of stochastic system control has become one of the hot spots in control research.Markov jump system is a special class of stochastic systems,used to describe the element interconnect fault or by sudden environment changes and other factors,system parameters and structure diversity changes,the law satisfies the Markov process.As the Markov jump system has a good practical application background,the control problem of Markov jump system has gradually become a new field of control research,and has been widely used in practice.Model Predictive Control?MPC?is considered to be an advanced process control strategy which can effectively deal with factors such as system constraints and parameter uncertainties.Robust model predictive control?RMPC?,as one of the branches of MPC,can improve the robustness of the system while ensuring the good control performance of the predictive control algorithm.In recent years,more and more attention has been paid to the study of robust predictive.In this paper,The problem of robust predictive control for a class of Markov jump systems is researched.The main contributions are described as follows:1.The problem of hybridH₂/H_?robust model predictive control for a class of Markov jump systems with constraints and bounded disturbance is researched.A hybridH₂/H_?performance indicators,at the same time,to ensure good control performance of system.In order to improve the control system of disturbance rejection ability and expand the initial feasible region,predictive controller based on multi-step control set is designed.Further guaranteed sufficient conditions of recursive feasibility,mean square stability of the closed-loop system is proved.2.The problem of robust distributed predictive control for a class of Markov jump systems with polyhedral uncertainties is researched.In order to effectively reduce the online computation and the computational complexity,the Markov jump system isdecomposed into several subsystems.At the same time,by using the dual-mode predictive control method,the mean square stability of the closed-loop system is guaranteed by the proposed dual-mode predictive controller.Finally,the simulation examples are given to demonstrate the effectiveness of the proposed control algorithm.