In this paper, the H ∞ control problem is investigated for a general class of discrete-time nonlinear stochastic systems with state-, control-, and disturbance-dependent noises (also called (x, u, v)-dependent noises). In the system under study, the system state, the control input, and the disturbance input are all coupled with white noises, and this gives rise to considerable difficulties in the stability and H ∞ performance analysis. By using the inequality techniques, a sufficient condition is established for the existence of the desired controller such that the closed-loop system is mean-square asymptotically stable and also satisfies H ∞ performance constraint for all nonzero exogenous disturbances under the zero-initial condition. The completing square technique is used to design the H ∞ controller with hope to reduce the resulting conservatism, and a special algebraic identity is employed to deal with the cross-terms induced by (x, u, v)-dependent noises. Several corollaries with simplified conditions are presented to facilitate the controller design. The effectiveness of the developed methods is demonstrated by two numerical examples with one concerning the multiplier-accelerator macroeconomic system.
883has been developed in the work of Lin and Byrnes 6 for discrete-time nonlinear deterministic systems. By means of matrix inequalities and Riccati equations, the stochastic bounded real lemmas have been proposed and the corresponding H ∞ control problem has been investigated in the works of El Bouhtouri et al 3 and Hinrichsen and Pritchard 4 for continuousand discrete-time linear stochastic systems, respectively.Since stochasticity and nonlinearity are two main sources contributing to the system complexities in practice, 7-17 the H ∞ control and filtering problems of nonlinear stochastic systems have been extensively investigated. In other works, 18-20 the nonlinear effects have been treated as external disturbances, satisfying certain conditions such as Lipschitz and sector-bounded conditions, and the proposed methodology might not be applicable to more general nonlinear systems. To deal with more general nonlinear stochastic systems, several sufficient conditions have been derived in the work of Zhang and Chen 20 for the H ∞ control problem by means of Hamilton-Jacobi equations/inequalities. For a large class of discrete-time nonlinear stochastic systems, a bounded real lemma has been established in the work of Berman and Shaked 21 for synthesizing a state feedback H ∞ controller. Recently, the quantized H ∞ control problem has been addressed in the work of Wang et al 22 for a class of nonlinear stochastic time-delay network-based systems with probabilistic data missing. To this end, it can be concluded that the mostly frequently used methodologies for designing H ∞ controllers for general nonlinear stochastic systems are second-order nonlinear Hamilton-Jacobi equations (HJEs) 20 and Hamilton-Jacobi inequalities (HJIs). 18,22 Traditionally, in the context of stochastic and/or H ∞ control, the disturbance...