Stochastic morphological evolution equations (Q640050)

The authors extend the concept of deterministic morphological evolution equations developed by \textit{J.-P. Aubin} [Mutational and morphological analysis. Tools for shape evolution and morphogenesis. Systems and Control: Foundations and Applications. Boston, MA: Birkhäuser (1999; Zbl 0923.58005); Set-Valued Anal. 1, No. 1, 3--46 (1993; Zbl 0784.34015)] to a stochastic setting of set-valued Itô stochastic evolution equations in separable Hilbert spaces. The purpose of the newly introduced equations is to overcome the difficulties of locally defined equations in modeling morphological evolution. The equations introduced here are stochastic integral equations written symbolically as differential equations. The existence of solutions and their dependence on initial data are established. The dynamics of this approach is not restricted to convex or expanding sets, allowing to obtain a more realistic treatment of nucleation and birth-and-growth processes in a random environment. The paper starts with a collection of definitions and properties that are relevant to random closed sets, set-valued maps and Itô processes. After that, the concept of a reachable set is extended to stochastic differential inclusions in a separable real Hilbert space and some of their properties are investigated. These sets are used as generalised integrators for the stochastic morphological evolution equations. The nonlocal equations generate a deterministic nonautonomous dynamical system formulated as a two-parameter semigroup. Their solutions model shape evolutions with focus on growth and translation. The last section shows how the model can handle the nucleation processes by means of approximation.