Solution of an abstract Cauchy problem with nonlinear and random perturbations in the Colombeau algebra (Q2375972)

The authors study the quasi-linear stochastic Cauchy problem \(X'_t=AX_t+F(X_t)+BW_t\), \(t\geq 0\), \(X_0=f\), where \(A\) is the generator of a \(C_0\)-semigroup of linear operators over a Hilbert space \(H\) or of an integrated semigroup, \(W\) is an \(H\)-valued time-white noise, \(F\) is a nonlinear mapping over \(H\) and \(B\in {\mathcal L}(H,H_a)\), where \(H_a\) is an algebra in \(H\). A way to study such an equation in the linear case (\(F=0\)) is to overcome the irregularity of the white noise \(W\) by studying the equation in spaces of abstract distributions. While the linear equations have been well understood by now, a translation of this approach to nonlinear equations involves multiplying generalised functions, which, in addition, are Hilbert space-valued. In their paper, the authors adapt Colombeau's approach for the multiplication of generalised functions and the construction of generalised solutions of the above equation. Their main result is established for the case \(H=L^2(\mathbb R)\), \(F\in C^\infty_b(\mathbb R)\) with \(F(0)=0\) and \(H_a\) is the set of \(m\) times differentiable functions in \(L^2(\mathbb R)\) belonging to the domain of \(A\). A typical example is that of a differential operator \(A\).