Article
Transactions of the Institute of
Measurement and Control
2014, Vol. 36(2) 283–288
Ó The Author(s) 2013
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0142331213501689
tim.sagepub.com
Finite-time stochastic stabilization for
uncertain Markov jump systems
subject to input constraint
Zhiqiang Zuo, Hongchao Li and Yijing Wang
Abstract
In this paper, the problem of finite-time stochastic stabilization for a class of uncertain Markov jump systems subject to partly known transition prob-
abilities and disturbances is investigated. The constraint on the control input sent to the actuator is also considered. A finite-time stochastic stabiliza-
tion criterion is given, to ensure that the state trajectory remains in a bounded region of the state space over a fixed time interval in mean square
sense. An optimization algorithm is proposed to find the optimal performance and the corresponding feedback controller where the undesirable high-
gain phenomenon can be avoided. Finally, numerical examples are given to illustrate the effectiveness of the proposed method.
Keywords
Finite-time stabilization, input constraint, Markov jump systems, uncertainty
Introduction
In some practical applications, the main concern is the beha-
vior of the system over a specified time interval. That is, given
some initial conditions, the state remains within a prescribed
bound over a fixed time interval: we call it finite-time (or
short-time) stable. The definition of finite-time stability was
firstly introduced in (Dorato, 1961). This kind of stability can
be used in many practical applications. For instance, the
problem of not exceeding the given bounds for the state tra-
jectories in a prescribed time interval, systems with saturation
elements in the control loop and so forth (Amato and Ariola,
2005; Amato et al., 2006). Moreover, the external distur-
bances are inevitable in practice. The problem of finite-time
control for systems subject to disturbances was studied in
(Amato et al., 2001).
Practical systems with abrupt changes and breakdowns of
components, such as manufacturing systems, communication
systems, power systems and aerospace systems can be repre-
sented by Markov jump systems. Krasovskii and Lidskii
(1961) built the formalism of Markov jump systems and stud-
ied the optimal control problem. Markov jump systems have
received a lot of attention in recent years (Shi et al., 1999;
Cao and Lam, 2000; Boukas, 2005; Karan et al., 2006; Dong
and Yang, 2008; Liang et al., 2009; Liu et al., 2009) due to its
great application value. However, the transition probabilities,
which determine the system’s behavior, are usually partly
known or costly to obtain. Therefore, many researchers have
studied this problem and some results have been reported for
Markov jump systems with partial information of transition
probabilities in recent years (Zhang and Boukas, 2009a,
2009b, 2009c; Zhang and Lam, 2010; Zhang et al., 2011).
Recently, the problem of finite-time stochastic stability and
stabilization for Markov jump systems has received increas-
ing interest. Zuo et al. (2012) studied the finite-time stability
and stabilization for Markov jump systems subject to partial
information on transition probabilities.
On the other hand, control systems are always subject to
input constraint because of the physical limitation. Hence, the
study of systems with input constraint is of considerable
importance. The input constraint has been well addressed in
many systems; see, for example, Du and Zhang (2009) and
Zhang and Evans (1994). However, to the best of our knowl-
edge, the finite-time stochastic stabilization for Markov jump
systems with input constraint has not been explicitly dealt with.
Motivated by the above discussion, we investigate the
finite-time stabilization for Markov jump systems with input
constraint in this paper. The system to be studied is more gen-
eral compared with the existing results. More specifically, the
transition probabilities are not completely known, and the
considered system involves norm- bounded uncertainties. A
criterion is formulated to guarantee that the system can be
finite-time stochastically stabilized while taking the input con-
straint into consideration simultaneously. Such a treatment
will avoid the undesirable high-gain feedback phenomenon.
In the final section, numerical examples are shown to demon-
strate the validity of the proposed result.
Tianjin University, PR China
Corresponding author:
Zhiqiang Zuo, Tianjin Key Laboratory of Process Measurement and
Control, School of Electrical Engineering and Automation, Tianjin,
300072, PR China.
Email: zqzuo@tju.edu.cn