Ivlodel Predictive Control Method For Markov Jump Systems

Model predictive control (MPC) can effectively handled the complex systems with hardmodeling and constraints. Markov jump systems are a class of complex systems, which mayexperience diversity variations in their parameters and structure, caused by phenomena ascomponent failures, or abrupt environmental changes, etc, and simultaneously have continuestime and discrete event two modes in the operational process. So, the research that modelpredictive control for a class of Markov jump systems is significative.The robust predictive controller designed is studied by multi-step strategy or efficientrobust control method for the discrete-time Markov jump systems. The main contents arestated as follows:1. With consideration of the discrete-time Markov jump systems with additive disturbance,the receding horizon H∞control problem is studied. A H∞predictive controller is designedbased on multi-step control strategy with partly unknown transition probabilities betweenjump modes, which results in the larger feasible region. The sufficient conditions of theensuring the recursive feasibility of robust model predictive control problem is investigated.2. A dual-mode control problem is discussed for a class of uncertain discrete-time Markovjump systems subject to input and state constraints. An efficient robust predictive controller isdesigned, which transfers the main online computations to be completely offline. Theproposed method can reduce computational burden by combining offline computations withonline optimization. The controller designed guarantees the closed-loop system mean squarestable.