Stein's method and point process approximation (Q1201756)

By using the Stein-Chen method an upper bound for the total variation \(d_{TV}({\mathcal L}(\Xi),Po(\lambda))\) between the distribution \({\mathcal L}(\Xi)\) of a simple point process \(\Xi\) on a compact, second countable Hausdorff space \(\Gamma\) and the distribution \(Po(\lambda)\) of a Poisson process on \(\Gamma\) with mean measure \(\lambda\) is proved. As an example this result is used to estimate the error when the uniform hard-core point process on the \(d\)-dimensional torus \(\Gamma\) is approximated by a homogeneous Poisson process.