The application of the separation principle for the linear continuous systems with coloured noise (Q805588)

The system \(x(t)\) satisfies (1) \(dx=Ax dt+Bu dt+G dW\) and the observation satisfies the following system \[ (2)\quad y=Cx+z\quad (3)\quad dz=Nz dt+R d V \] where A,B,G are matrices which usually depend on time t,C,N,R are real matrices and V,W are independent brownian notions adapted to the same increasing family of \(\sigma\) algebras \((F_ t)\). The process \(z=z(t)\) is the coloured noise defined by (3). First using the Kalman- Bucy filter model the author gets the mean square optimal estimate of x, \(\hat x(t)\). Then he shows that the separation principle is applicable for the optimal control problem with cost \[ (4)\quad J(u)=E(\int^{T}_{0}L(t,x(t),u(t)dt+\psi (x(T))) \] and finally gives the explicit form of the optimal control in the scalar autonomous case.