A numerical scheme for backward doubly stochastic differential equations (Q1940750)

For the class of backward doubly stochastic differential equations (BDSDEs) with two independent Brownian motions \(B\) and \(W\) which have possible path-dependent values the author derives a numerical scheme. First, the \(L^2\)-regularity of the martingale integrand \(Z\) under some conditions is shown. Then, the solution \((X,Y,Z)\) of a BDSDE whose terminal value has a special form is approximated by using ``step processes''. The approximation scheme consists of conditional expectation w.r.t. the enlarged filtration. Moreover, the backward component and the associated control \((Y,Z)\) can be expressed as a function of \(X\) and \(B\). As an application, it can be used for determining option prices on financial assets whose dynamics are solutions of SPDEs.