On curious properties of the general size-biased distribution
Alexander E. Patkowski
TL;DR
The paper studies a general size-biased distribution associated with the density $v_k(x)$, obtained from a Mellin-type transform of the Riemann $\xi$-function, and proves $v_k(x)>0$ for $k\ge1$ along with a majorization $X_{k-1}$ majorizing $X_k$. It develops a functional-equation framework that yields an identity linking moments of $X_k$ to $\xi^k(y)$ and expresses a convolution relation $v_{m+k}(z)=\int_0^{\infty} v_k(x) v_m(xz) dx$. It connects to the Riemann hypothesis by showing RH is equivalent to the real-zero property of $\Xi_0(t)$, and extends to a family $\Xi_{\lambda,k}(t)$ that obeys a backward-heat equation, widening zero-distribution criteria. Overall, the work ties analytic number theory and probabilistic constructions to offer new insights into xi-function related transforms via size-biased random variables.
Abstract
We offer further results on a general size-biased distribution related to the Riemann xi-function we presented in [9] using the work of Ferrar. Curious properties associated with its expected value are presented, which are related to special functional equations. We also relate our observations to some recent developments related to the Riemann hypothesis.
