A solvable class of renewal processes (Q829352)

When the interarrival times in a discrete renewal process have a geometric distribution \(K(n)=x(1-x)^{n-1},n\geq 1\), then the resulting renewal process is simple and fully known. In the paper the authors consider mixtures of the form \(K(n)=\int_{0}^{1}(1-x)^{n-1}xd\mu (x)\), \(n\geq 1\) and \(K(\infty )=\mu (\left\{ 0\right\} )\). Using Stieltjes transforms they prove that the renewal sequence is given by \(u(n)=\int_{0}^{1}x^{n}d\nu (x)\) for a specific probability measure \(\nu \) related to \(\mu \). In the continuous case the authors consider exponential mixtures \(f(x)=\int_{0}^{\infty }se^{-sx}d\mu (s)\) and find that the renewal intensity is given by \(U(x)=\int_{0}^{\infty }e^{-sx}d\nu (s)\) for a positive measure \(\nu \) related to \(\mu \). As an application the authors consider a random polymer model and obtain explicit formulas in the case of geometric mixtures.