Approximate solvability of forward-backward stochastic differential equations (Q5956450)

Let \(W\) denote a \(d\)-dimensional Brownian motion. An triplet of square integrable processes \(\Sigma= (X,Y,Z)\) adapted w.r.t. the filtration generated by \(W\) is called a solution to the forward-backward stochastic differential equation (FBSDE) with initial condition \(x\in \mathbb{R}^n\) and coefficients \(a\), \(\sigma\), \(h\), \(g\) if \[ (1)\qquad dX_t= b(t,\Sigma_t) dt+ \sigma(t,\Sigma_t) dW_t,\qquad (2)\qquad X_0= x,\qquad (3)\qquad Y_T= g(X_T). \] FBSDEs generalize the backward SDEs studied by \textit{E. Pardoux} and \textit{S. G. Peng} [Syst. Control Lett. 14, No. 1, 55-61 (1990; Zbl 0692.93064)]. Since the first paper on FBSDEs by \textit{F. Antonelli} [Ann. Appl. Probab. 3, No. 3, 777-793 (1993; Zbl 0780.60058)], there have been a lot of works in this subjet, among them the papers of the authors [Chin. Ann. Math., Ser. B 16, No. 3, 279-298 (1995; Zbl 0832.60067)], the authors and \textit{P. Protter} [Probab. Theory Relat. Fields 98, No. 3, 339-359 (1994; Zbl 0794.60056)], \textit{Y. Hu} and \textit{S. Peng} [ibid. 103, No. 2, 273-283 (1995; Zbl 0831.60065)] and \textit{S. Peng} and \textit{Z. Wu} [SIAM J. Control Optimization 37, No. 3, 825-843 (1999; Zbl 0931.60048)]. These papers present different methods to solve FBSDE (1)--(3); the type of assumptions on the coefficients they use are different, but they have in common that they are rather restrictive. Working only with rather mild assumptions, the authors study the approximate solvability of the FSBE, by which is meant that, for any \(\varepsilon> 0\) there is some \(\Sigma\) satisfying (1), (2), and \((3')\) \(E[|Y_T- g(X_T)|]\leq \varepsilon\). In their work they develop a general scheme for constructing approximate solutions.