A general class of bounded diffusion processes with applications in the biosciences

This paper develops and analyzes a unified class of bounded continuous-time diffusion processes defined by stochastic differential equations. The models share a common random component but differ in their drift specification, encompassing polynomial drift of arbitrary order as well as generalized logistic and Gompertz growth dynamics. The polynomial drift class unifies and extends several widely used bounded diffusions in the biosciences, including stochastic logistic growth, Jacobi processes and Wright-Fisher models. For all model classes, we establish existence results and provide a complete classification of boundary behavior, yielding transparent conditions for the existence of stationary distributions. The diffusion specification permits both symmetric and asymmetric behavior near the boundaries. Whenever a stationary distribution exists, explicit expressions for its density (up to a normalizing constant) are derived. The resulting framework substantially expands the set of tractable bounded diffusions available for empirical modeling of bounded biological processes.