Stationary distribution of a two-dimensional SRBM: geometric views and boundary measures (Q1955513)

The paper considers a two-dimensional semimartingale-reflecting Brownian motion (SRBM) in the non-negative quadrant. The data for SRBM consists of a covariance matrix \(\Sigma \), a drift vector \(\mu \), and a reflection matrix \(R\). The covariance matrix and drift vector determine the ellipse \(\frac{1}{2}(\theta ,\Sigma \theta ) + (\mu ,\theta ) = 0\). In addition to this ellipse, two lines play an essential role in the paper. They pass through the origin and are orthogonal to the first and to the second column of the reflection matrix, respectively. The paper presents three sets of results for SRBM. It gives a geometric interpretation for the value function of the variational problem that is associated with the SRBM, a geometric condition for the SRBM to have a product form stationary distribution, and presents exact tail asymptotics of two boundary measures that are associated with the stationary distribution.