The Arov-Grossman model and the Burg multivariate entropy (Q1879335)

The article deals with the Arov-Grossman functional model to describe all the solutions of the covariance extension problem for \(q\)-variate stationary stochastic processes. The main problem is to find the maximum of the functional \(\varepsilon(F)= \frac1{2\pi}\int^{2\pi}_0 \log\det F(e^{it})\,dt\) subject to the conditions \(\frac1{2\pi} \int^{2\pi}_0 e^{-ikt}F(e^{it})\,dt =R_k,\) \( |k|\leq p,\) where the coefficients \(\{R_k\}^p_{k=-p}\) are the covariances of the stationary \(q\)-variate stochastic process.