Refined distributional approximations for the uncovered set in the Johnson-Mehl model. (Q1766018)

Whereas the main volume of work concerning Johnson-Mehl model relates to the properties of the corresponding tessellation (the relevant references are summarized in the introduction), the author systematically devotes his attention to the properties of the uncovered set [Stochastic Processes Appl. 65, 31--53 (1996; Zbl 0889.60054)]. The result of the previous paper [Probab. Theory Relat. Fields 117, 145--161 (2000; Zbl 0961.60084)] on stationary reward processes is used here to derive a bound for the total variation distance between the distribution of the number of components of the uncovered set at time \(z\) intersecting \([0,t]\) and a compound Poisson-geometric distribution. The bound is simpler and sharper than that found in [J. Appl. Probab. 37, 101--117 (2000; Zbl 0958.60008)]. As a completely new result, a bound is given for the total variation distance between the distribution of the Lebesgue measure of the part of the uncovered set at time \(z\) and a compound Poisson-exponential distribution. The two parameters of the approximating distributions are explicitly calculated in the both cases. Finally, a non-uniform error bound is obtained for the approximation of the distribution of the time of complete tessellation by an extreme value Gumbel distribution.