Receding horizon H_∞ control for discrete-time Markovian jump linear systems

Receding horizon H∞ control scheme which can deal with both the H∞ disturbance attenuation and mean square stability is proposed for a class of discrete-time Markovian jump linear systems when minimizing a given quadratic performance criteria. First, a control law is established for jump systems based on pontryagin’s minimum principle and it can be constructed through numerical solution of iterative equations. The aim of this control strategy is to obtain an optimal control which can minimize the cost function under the worst disturbance at every sampling time. Due to the difficulty of the assurance of stability, then the above mentioned approach is improved by determining terminal weighting matrix which satisfies cost monotonicity condition. The control move which is calculated by using this type of terminal weighting matrix as boundary condition naturally guarantees the mean square stability of the closed-loop system. A sufficient condition for the existence of the terminal weighting matrix is presented in linear matrix inequality （LMI） form which can be solved efficiently by available software toolbox. Finally, a numerical example is given to illustrate the feasibility and effectiveness of the proposed method.

Reaction Subdiffusion with Random Waiting Time Depending on the Preceding Jump Length

To describe the energy-dependent characteristics of the reaction-subdiffusion process, we analyze the simple reaction A--→B under subdiffsion with waiting time depending on the preceding jump length, and derive the corresponding master equations in the Fourier Laplace space for the distribution of A and B particles in a continuous time random walk scheme. Moreover, the generalizations of the reaction-diffusion equation for the Gaussian jump length with the probability density function of waiting time being quadratically dependent on the preceding jump length are obtained by applying the derived master equations.

Real-Time Detection for Magnetic Island of Neoclassical Tearing Mode in EAST Plasma Control System

Accurate detection of a magnetic island in real time is one of the important issues for the tearing mode（TM） and neoclassical tearing mode（NTM） control.This paper presents a real-time detection system for the magnetic island of NTM control in the EAST Plasma Control System（PCS）.Diagnosis is based on magnetic periodic perturbation and electron temperature fluctuation caused by the magnetic island.Therefore,a Mirnov measurement has been selected to calculate the island＇s parameters,such as island width,frequency of island rotation,and toroidal number.The electron cyclotron emission（ECE） system can detect the island position,which is calculated by two fast detection algorithms called correlation analysis and Hilbert transform.For future NTM control,real-time equilibrium reconstruction（rt-EFIT） is needed to locate the rational q-surface where the island is detected.This fast detection system is able to detect an island within 3 ms.It can be integrated into PCS to provide effective parameters of the island for NTM control by using EC resonance heating（ECRH） in the next experiment of EAST.