Modern control systems are generally composed of interacting subsystems with both continuous and discrete dynamics. In particular, the investment portfolio is a complex system, which can be consisted of risky financial assets of different classes, where the dynamics of returns varies discontinuously in accordance with the evolution of interrelated states of Markov chains describing, for example, the behavior of the various sectors of economy. In this paper we consider complex Markov jump linear system composed of interconnected subsystems. The parameters of each subsystem change in accordance with the evolution of the simple connected Markov chains whose states are interrelated. We use model predictive control approach to solve the problem. The open-loop feedback control strategy is derived taking into account explicit constraints on the input variables. Predictive strategies computation includes the decision of the sequence of quadratic programming tasks.

В работе рассматривается задача синтеза оптимальных стратегий управления с прогнозирующей моделью для гибридных систем, состоящих из подсистем, с учетом ограничений на управляющие переменные. Параметры каждой из подсистем изменяются в соответствии с эволюцией марковских цепей, состояния которых взаимосвязаны между собой.