Coupling all the Lévy stochastic areas of multidimensional Brownian motion (Q2370093)

Let \(A_t\), \(B_t\) be two \(n\)-dimensional Brownian motions adapted to the same filtration (co-adapted\-ness). It is shown that a successful coupling of \(A\), \(B\) and their respective stochastic area processes \(\int A_i\,dA_j-\int A_j \,dA_i\) and \(\int B_i \,dB_j-\int B_j\,dB_i\), \(i,j= 1,2,\dots, n\) is possible and the coupling is explicitly constructed. The main theorem gives conditions under which this coupling is possible in finite time with positive probability, resp., almost surely. Technically, co-adaptedness allows the usage of Itô's stochastic calculus and an ansatz of the form \(dA= J^T dB+\widetilde J^T dC\) -- with \(J\), \(\widetilde J\) being predictable \(\mathbb{R}^{n\times n}\)-valued processes and \(C_t\) being a further \(n\)-dimensional Brownian motion, co-adapted and independent of \(B\) -- turns out to give a coupling of \(A\) and \(B\). Of course, additional restrictions are needed to get a simultaneous coupling of the Lévy areas. At the end of the paper it is conjectured that one can replace the Lévy areas by a finite set of iterated Wiener integrals, subject to algebraic compatibility of the initial conditions.