The branching process with logistic growth

coalescence process continuous-state branching process density dependence fragmentation logistic process Ornstein-Uhlenbeck type process population dynamics Riccati differential equation Size-dependent branching process

Mathematics Subject Classification ID

Branching processes (Galton-Watson, birth-and-death, etc.) (60J80)

Abstract: In order to model random density-dependence in population dynamics, we construct the random analogue of the well-known logistic process in the branching process' framework. This density-dependence corresponds to intraspecific competition pressure, which is ubiquitous in ecology, and translates mathematically into a quadratic death rate. The logistic branching process, or LB-process, can thus be seen as (the mass of) a fragmentation process (corresponding to the branching mechanism) combined with constant coagulation rate (the death rate is proportional to the number of possible coalescing pairs). In the continuous state-space setting, the LB-process is a time-changed (in Lamperti's fashion) Ornstein-Uhlenbeck type process. We obtain similar results for both constructions: when natural deaths do not occur, the LB-process converges to a specified distribution; otherwise, it goes extinct a.s. In the latter case, we provide the expectation and the Laplace transform of the absorption time, as a functional of the solution of a Riccati differential equation. We also show that the quadratic regulatory term allows the LB-process to start at infinity, despite the fact that births occur infinitely often as the initial state goes to infty. This result can be viewed as an extension of the pure-death process starting from infinity associated to Kingman's coalescent, when some independent fragmentation is added.

Review text

A stochastic analogue of the deterministic logistic growth model is constructed: Let \(X\) be a spectrally positive Lévy process with exponent \[ \psi(\lambda)= \alpha\lambda+ (\gamma/2)\lambda^2+ \int^\infty_0 (e^{-\lambda r}- 1+\lambda r\mathbf{1}_{r< 1})\Pi(dr),\quad \lambda\geq 0, \] where \(\alpha\), \(\gamma\) are real numbers, \(\gamma> 0\), and \(\Pi\) is a positive \(\sigma\)-finite measure satisfying \(\int^\infty_0 (1\wedge r^2)\Pi(dr)< \infty.\) Let \(c\) be a real number, \(R\) the unique strong solution of \(dR(t)= dX_t- cR(t)\,dt\), \(t> 0\), \(R_0= 0\), \(\tau\) the finite hitting time of \(0\), and \(C_t\) the right-inverse of \(\eta_t= \int^{t\wedge\tau}_0(1/R(s))\,ds\), \(t> 0\). The logistic branching process \(Z\) is then defined as the Feller process \(Z_t:= R(C_t)\) for \(0\leq t<\eta_\infty\), \(Z_t:= 0\) for \(\eta_\infty< \infty\wedge t\geq \eta_\infty\). The infinitesimal generator of \(Z\) is \(xA- cx^2d/dx\), where \(A\) is the infinitesimal generator of \(X\). If \(Z\) has only continuous paths, then it is a Feller diffusion with logistic growth, \(\psi(\lambda)= (\gamma/2)\lambda^2- b\lambda\), \(dZ_t= Z_t dt- cZ^2_t dt+ (\gamma Z_t)^{1/2} dB_t\), \(t> 0\), \(B\) standard Brownian motion. If \(\psi\) is the exponent of an increasing Lévy process or subordinator (no ``natural deaths), then \(Z\) is recurrent. If \(Z\) has zero drift (no ``infinitesimal births) and Lévy measure of finite mass \(\rho< c\), then \(Z\) is null-recurrent and \(Z_t\to 0\) in probability. Otherwise, \(Z\) is positive-recurrent. When there are no ``natural deaths, then \(Z_t\to 0\) almost surely. The integer-valued analogue of \(Z\) is also defined. Its essential properties are properties of the real-valued process, but not vice versa. The sequence of processes obtained by suitably scaling the integer-valued process converges to a Feller process with logistic growth.

Recommendations

Cites work

Cited in

(73)

This page was built for publication: The branching process with logistic growth

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q558688)