A note on large deviations for the stable marriage of Poisson and Lebesgue with random appetites (Q2428529)

Let \(\Xi\subset{\mathbb R}^d\) be a set of centers chosen according to a Poisson point process of law \(\text{P}\) and expectation \(\text{E}\), whose intensity is \(\lambda=\operatorname{E}[\Xi\cap[0,1]^d]\). Let \(\alpha\) be a nonnegative random variable with law \({\mathcal P}\) and expectation \({\mathcal E}\) called appetite. Every center \(\xi\in \Xi\) has a random appetite \(\alpha_{\xi}\) in such a way that appetites corresponding to different centers are independent of each other and identically distributed. The appetites are also independent of \(\Xi\). The elements \(x\in {\mathbb R}^d\) are called sites. Let \(\Psi\) be a function of \({\mathbb R}^d\) to \(\Xi\cup\{\infty, \Delta\}\) called allocation. The goal of \(\Psi\) is to allocate every site to a center. If \(\Psi\) cannot warrant that allocation, then the site is sent to the element \(\infty\). Also, there is a null set of sites under the Lebesgue measure \({\mathcal L}\) which will be sent to \(\Delta\). In [\textit{C. Hoffman, A. Holroyd} and \textit{Y. Peres}, ``A stable marriage of Poisson and Lebesgue'', Ann. Probab. 34, No. 4, 1241--1272 (2006; Zbl 1111.60008)], an explicit construction of the function \(\Psi\) for the case of constant appetite for every center is given, and it was extended for the general case of random appetites in [\textit{D. Diaz}, ``Algumas propriedades de alocações pontual de Poisson'', Ph.D. thesis, Instituto de Matemática e Estatística, Cidade Universitária, São Paulo (IME-USP), Brasil (2009)]. The authors of the present paper apply a construction for the Poisson Boolean model by \textit{R. Meester} and \textit{R. Roy} [Continuum percolation. Cambridge: Cambridge University Press (1996; Zbl 0858.60092)]. Generalizing previous results, the authors study large deviations for the distance of a typical point \(x\in {\mathbb R}^d\) to its center \(\Psi(x)\in\Xi\), subject to some restrictions on the moments of \(\alpha\).