New results in Branching processes using Stochastic Approximation

TL;DR

This paper analyzes complex branching processes using stochastic approximation, demonstrating convergence properties and extending existing models to new variants, with applications in viral markets and social networks.

Contribution

It introduces new variants of branching processes and applies stochastic approximation to analyze their long-term behavior, extending prior theoretical frameworks.

Findings

Proportion of populations converges to equilibrium points or nears saddle points.

Normalized trajectories of the embedded chain converge to ODE solutions.

New variants of branching processes with attack, acquisition, and proportion-dependent offspring are analyzed.

Abstract

We consider a broad class of continuous-time two-type population size-dependent Markov Branching Processes. The offspring distribution can depend on the current (alive) and total (dead and alive) populations. Using stochastic approximation techniques, we show that the time-asymptotic proportion of the populations either converges to the equilibrium points or infinitely often enters every neighbourhood and exits some neighbourhood of a saddle point of an appropriate ordinary differential equation with a certain probability (almost surely for the process with attack and proportion-dependent branching process). The result holds under finite second-moment conditions. We also show that certain normalized trajectories of the embedded chain almost surely converge to the solution of the ordinary differential equation uniformly over any finite time window as time progresses. In addition to extending the analysis of several existing BPs, we analyze two new variants: BP with attack and acquisition, and BP with proportion-dependent offspring. Using these results, we study competition in viral markets and fake news control on online social networks.