Reflected forward-backward SDEs and obstacle problems with boundary conditions (Q5950196)

The authors prove the existence and uniqueness of the solution \(\Theta= (X, Y,Z)\) of the reflected forward-backward SDE \[ dX_t= b(t,\Theta_t) dt+ \sigma(t,\Theta_t) dW_t+ d\eta_t,\;dY_t= h(t,\Theta_t) dt- Z_t dW_t+ d\zeta,\quad t\in [0,T], \] \[ X_0= x,\;Y_T= g(X_T), \] where \(W\) is the driving Brownian motion and \(\eta\), \(\zeta\) are adapted bounded variation processes keeping \(X\) and \(Y\) from leaving certain prescribed regions \(K\) and \(A\), respectively; \(K\) is supposed to be a deterministic convex while the convex \(A\) can be random. The proof uses the method of contraction mapping, in the spirit of \textit{E. Pardoux} and \textit{S. Tang} [Probab. Theory Relat. Fields 114, No. 2, 123-150 (1999; Zbl 0943.60057)], with some modifications. In the second part of the paper the authors show that in the Markovian framework the solution \(Y\) gives the stochastic interpretation of the unique viscosity solution to the associated variational inequality with Neuman boundary. A small paragraph of the paper is devoted to applications to an American game option. The authors' paper is a precious contribution to the subject of backward SDEs and forward-backward SDEs [pioneer works by \textit{E. Pardoux} and \textit{S. G. Peng}, Syst. Control Lett. 14, No. 1, 55-61 (1990; Zbl 0692.93064) and \textit{F. Antonelli}, Ann. Appl. Probab. 3, No. 3, 777-793 (1993; Zbl 0780.60058)]. The reader is also referred to recent works on backward SDEs with reflections [\textit{N. El Karoui}, \textit{C. Kapoudjian}, \textit{E. Pardoux}, \textit{S. Peng} and \textit{M. C. Quenez}, Ann. Probab. 25, No. 2, 702-737 (1997; Zbl 0899.60047); \textit{J. Cvitanić} and \textit{I. Karatzas}, ibid. 24, No. 4, 2024-2056 (1996; Zbl 0876.60031)].