The design of large stochastic optimal regulating systems is considered. The system model represents a typical industrial-control situation, with coloured noise and disturbance inputs to the plant. The resulting system involves a large number of state variables, and it is inefficient to solve the control and filtering Riccati differential equations for this system. It is shown that the redundancy in the system equations can be used to reduce the order of the Riccati equations which must be solved. It is also shown that a steady-state (or constant-gains) solution to the problem exists, and is unique. This system is also shown to be asymptotically stable.In the design of industrial controllers, it is often necessary to include integral action to offset constant disturbances. This is also desirable in the above situation, even when most of the disturbances are modelled accurately, if the regulating error is to be reduced to zero in the steady state. The integral control is introduced by including an additional term, which involves an integral operator, in the usual cost function. The performance criterion is also specified, so that the system regulates about given set-point values. The system is described in general terms, but the problem is based upon the dynamic ship-positioning control problem. The paper thus provides a formal solution to this problem, based upon l.q.g. stochastic optimal-control theory.