Distributed Receding Horizon Control For A Class Of Large Systems With Constraints

Large scale system refers to a system which is composed of several subsystems and widely exists in many fields,such as petroleum,chemical industry,electric power,transportation,etc.because of its large scale,strong coupling,uncertain constraint parameters and other characteristics,the centralized control cannot be implemented,The design idea of distributed control combined with centralized control and decentralized control can decompose the control task of complex large-scale system into multiple subsystems for control respectively and consider the coupling effect between each subsystem,which can reduce the complexity of control system and ensure the performance of control system,It is an effective strategy to solve the control task of uncertain large-scale system.Receding horizon control strategy is a brand-new system control strategy which flourished in the control practice of production industry in the 1970 s.Because of the unique rolling optimization and feedback correction and other basic elements of receding horizon control strategy,it can effectively deal with the control problem of large-scale system with complex and multiple constraintsIn the process of system control,system most of the information transmission through the network,inevitable phenomenon of the existence of constraint,the constraint can weaken the control performance of the system to a certain extent,will affect the stability of the system.This paper consider a class of linear constraints on global average,from two angles of discrete and continuous systems,to explore with average global linear constraints distributed receding horizon control stabilization problem for a class of large-scale systems.For discrete time large-scale systems with more complex constraints,by constructing a linear discrete forms of constraint model on average,will be constrained phenomenon caused by the nonlinear problem into a linear problem is studied thoroughly.The first use of distributed RHC method to deal with complex constraint on-line system characteristics and the properties of coupled differential Riccati equation,on the general objective function increase with constant coefficient of two terminal weighted matrix receding horizon control forms of the objective function,and establish a Lyapunov function,Obtained with the constraint of discrete time sufficientand necessary conditions of large system distributed receding horizon control stabilization.Secondly,for the necessary and sufficient condition of the system stabilization,are also given from the view of the necessary and sufficient proof.Finally,through the Pontryagin maximum principle is necessary condition is derived by two mutually coupled riccati equations,thus obtained can make for constraints discrete time large scale system stable explicit optimal distributed receding horizon control controller.For continuous time distributed control system with more complicated constraints,through the establishment of the linear continuous form of the average constraint model,using the method of distributed receding horizon control to discuss its decentralized stabilization problem of receding horizon control.According to the average of linear constraint model is established to analyze system performance index and the performance index increase two terminal weighted matrix,using predictive control comprehensive monotonicity of Lyapunov stability method,Lyapunov equation was established based on distributed receding horizon control forms of performance indicators,A set of Lyapunov function inequalities with two terminal weighted matrices are derived,and it is found that these two Lyapunov function inequalities are the necessary and sufficient conditions for the stabilization of distributed receding horizon control in large scale continuous time systems with constraints.