On non-Markovian forward-backward SDEs and backward stochastic PDEs (Q713213)

Let \((\Omega,{\mathcal F},\operatorname{P},{\mathbf F})\) be a filtered probability space on which is defined a \(d\)-dimensional Brownian motion \(B=(B_{t})_{t\geq0}\), \({\mathbf F}=({\mathcal F}_{t}^{B})_{t\geq0}\) be the natural filtration generated by \(B\), augmented by the \(\operatorname{P}\)-null sets of \({\mathcal F}\). The authors consider the following forward-backward stochastic differential equation \[ \begin{aligned} X_{t}&=x+\int_{0}^{t}b(s,X_{s},Y_{s},Z_{s})ds+\int_{0}^{t}\sigma(s,X_{s},Y_{s})dB_{s};\\ Y_{t}&=g(X_{T})+\int_{t}^{T}f(s,X_{s},Y_{s},Z_{s})ds-\int_{t}^{T}Z_{s}dB_{s}.\end{aligned} \] Here, \(X\in\mathbb{R}^{n}\), \(Y\in ^{m}\), \(Z\in\mathbb{R}^{m\times d}\), the coefficients \(b,\sigma, f, g\) can be random. It is supposed that (a) the coefficients \(b,\sigma, f\) are \({\mathbf F}\)-progressively measurable for any fixed \((x,y,z)\) and \(g\) is \({\mathcal F}_{T}\)-measurable for any fixed \(x\); (b)\ \(b,\sigma, f, g\) are uniformly Lipschitz in \((x,y,z)\); c)\ \(b\) and \(\sigma\) are bounded and \(\operatorname{E}\left[\int_{0}^{T}| f(t,0,0,0)|^2dt+| g(0)|^2\right]<\infty\). The authors introduce the notion of the decoupling random field, and establish the connection between the existence of the decoupling field and the well-posedness of the considered forward-backward stochastic differential equation. In the case \(b=b(t,\omega,x,y)\), \((t,\omega,x,y)\in [0,T]\times\Omega\times \mathbb{R}^{n}\times \mathbb{R}^{m}\), necessary and sufficient conditions in order for a random field to be a regular decoupling field of the considered forward-backward stochastic differential equation are obtained. Some properties of the backward stochastic partial differential equations, such as comparison theorem and stability are discussed.