Equilibrium and extinction in stochastic population dynamics (Q1099808)

The paper describes properties of stochastic population systems defined by a Markovian birth and death process. For large population sizes the system can be well approximated by a diffusion process. The corresponding forward equation contains a small parameter and can be asymptotically solved in the quasi-stationary state by the ray method. The deterministic system associated to the stochastic one is assumed to have a stable fixed point in the interior of the state space. In the first part of the paper a local analysis near this equilibrium is given and confidence regions are estimated by a new method of solving numerically the ray equations. The concept of extinction time which is closely related to the notion of stochastic stability is discussed. In the second part the theory is applied to a stochastic two-population model associated to a generalized Lotka-Volterra system. Stability conditions are derived and the parameters of the Gaussian distribution approximating the stationary density are calculated. Contours of confidence regions and rays are constructed by the numerical method developed by the author. Exit points and exit times are compared with simulations of the original birth and death process.