Receding Horizon Control For Several Kinds Of Stochastic Systems

Several kinds of stochastic systems are studied under the receding horizon control(RHC)scheme in this article.These include continuous-time large-scale systems with imposed constraints,continuous-time mean-field system,and stochastic discrete-time systems with one-step-delay sharing information structure.They are widely used in the real world,such as wind farms and communication networks,etc.So,it makes sense to study these three types of systems.The works are as follows:For continuous-time large-scale systems with constraints,their stabilization is analyzed using the RHC strategy.By using the variational method to transform the constraint problem,the explicit RHC controller is obtained by using the Pontryagin’s extremum principle.The stabilization conditions of the system are deduced under this RHC controller,namely,two inequality conditions,and the two inequalities contain terminal weighting matrices,then the RHC can guarantee the stability of system.The RHC stabilization problem is investigated for continuous-time mean-field systems without/with input delay.For the mean-field system without input delay,by constructing the performance index with two terminal matrices,solving the forward-backward stochastic differential equations(FBSDEs),we obtains an explicit RHC controller.And inequality conditions for system stability are obtained.On this condition,the system controlled by RHC is stable.For the mean-field system with input delay,the performance index containing more terminal matrices are constructed.The RHC controller and the stabilization inequality conditions of system are solved by the same method.For stochastic discrete-time system with one-step-delay sharing information structure,the stability is analyzed by RHC strategy.Firstly,by analyzing the information used by the two controllers,the system equation and performance index are rewritten.On this basis,the estimator is given based on the sharing information.Further,the RHC controller is solved by using the extremum principle.Finally,the condition of mean square bounded stabilization is obtained by using the RHC strategy and the boundness of error covariance.