zbMATH — the first resource for mathematics

Linear matrix inequalities, Riccati equations, and indefinite stochastic linear quadratic controls. (English) Zbl 0981.93080
The paper is concerned with an optimal stochastic LQ control problem in infinite time horizon. In contrast to the deterministic case, the control and state weighting matrices in the cost functional are indefinite. This leads to an indefinite LQ problem which may be well-posed due to the nature of uncertainty involved. The problem gives rise to a stochastic algebraic Riccati equation (SARE) which is different from the classical algebraic Riccati equation. To analyze the SARE, some linear matrix inequalities (LMI’s) are introduced whose feasibility is shown to be equivalent to the solvability of the SARE. A computational approach to the SARE is developed via a semidefinite programming associated with the LMI’s. Finally, numerical experiments are reported to illustrate the proposed approach.

93E20 Optimal stochastic control
49N10 Linear-quadratic optimal control problems
93B40 Computational methods in systems theory (MSC2010)
15A39 Linear inequalities of matrices
90C22 Semidefinite programming
Full Text: DOI