Random linear-quadratic mathematical models: Computing explicit solutions and applications (Q1013143): Difference between revisions

This paper deals with the study of two random population models. The first model is the linear random population differential model \[ \dot{X} (t) = a(t) X(t), \quad X(t_0) = X_0, \quad t_0, t \in \mathbb{R} \tag{1} \] where \(W_0\) is a random variable such that \(E W_0^2 < \infty\) and \(a(t)\) is a almost surely continuous and \(E(a(t))^2< \infty\) \(a(t)\) is independent of \(X_0\) for each \(t\in R.\) The authors prove that under some assumption the solution process of the model (1) is \[ X(t) = \exp \biggl\{ \int_{t_0}^t a(s)\,ds\biggr\} X_0 \] and find \(E X(t)\) and \(Var X(t).\) The particular case of (1) \(a(t) = m + \sigma B(t)\) is considered, \(B(t)\) is a Brownian motion and \(\sigma >0.\) The solution process process in this case is given by \[ X(t) = e^{mt} \exp \biggl\{ \sigma \int_0^t B(s)\,ds\biggr\} X_0. \] The second model is the random logistic population model \[ \dot{X} (t) = X(t)(a - bX(t)), \quad X(t_0) = X_0, \quad t \geq t_0 \] where \(a\) is the random reproductive parameter, \(b =a/k\) where \(k\) is the random carrying capacity, and \(X_0\) is the random initial population such that \(a, k, X_0\) are mutually independent. The authors prove that under some assumptions the solution process of this model is \[ X(t) = {a X_0 e^{a(t-t_0)}\over a - b X_0 + b X_0 e^{a(t -t_0)}}, \quad t \geq t_0. \] An example is considered as well.