Branching Brownian motion with absorption and the all-time minimum of branching Brownian motion with drift (Q2363162): Difference between revisions

In this paper, the authors study a dyadic branching Brownian motion on the real line with absorption at 0, drift \(\mu\in\mathbb{R}\) and start from a single particle at position \(x>0\). This means that particles move according Brownian motion with drift \(\mu\) and split into two particles at rate \(\beta\) independently one from another. Let \(K(t)\) be the number of particles that have been absorbed up to time \(t\). By \(K(\infty)\in\mathbb{N}\cup\{ \infty\}\) is denoted the limit \[ K(\infty)\, :=\, \lim\limits_{t\to\infty}K(t). \] The main results of the paper concern the study of \(K(\infty)\) in the regime \(\mu\geq\sqrt{2\beta}\). In this regime, \(K(\infty)\) is almost surely finite and for \(s\geq 0\) and \(x\geq 0\) \[ \omega_{s}(x)\, :=\, \mathbb{E}^x[s^{K(\infty)}],\qquad \omega_s(0)\, =\, s \] is introduced. It is shown that there exists a finite \(s_0>1\) depending only on \(\mu/\sqrt{\beta}\) such that for \(s\leq s_0\), \(\omega_s(x)\) is finite for all \(x\geq 0\), for \(s>s_0\), \(\omega_s(x)\) is infinite for all \(x>0\). The properties of \(\omega_s(x)\) and \(s_0\) are studied. Furthermore, the authors present three largely independent ways to characterize \[ \omega(x)\, :=\, \omega_0(x)\, =\, \mathbb{P}^x(K(\infty)=0). \] The first way is related to track the probability that no particle has been absorbed up to time \(t\) and used the connection between branching Brownian motion and the Kolmogorov-Petrovskii-Piskunov travelling wave equation, the second gives \(\omega(x)\) as the expectation of a certain martingale and the third one gives \(\omega(x)\) as a power series.