No label defined (Q4357500)

The authors prove a theorem on existence and uniqueness of solution for backward stochastic differential equations (BSDE) driven by a general càdlàg martingale. A solution of a BSDE is a triple \((Y,Z,N)\), where \(Y\) is an adapted càdlàg process, \(Z\) is a predictable process and \(N\) is a càdlàg martingale. Usually, the characteristic \(f\) of BSDE is supposed to be uniformly Lipschitz; to avoid this condition the authors look for solutions in a ``weighted space'', \(f\) satisfying non-uniform Lipschitz conditions. Namely, let \[ Y_t= \xi+ \int^T_t f(s, Y_s,Z_s)dC_s- \int^T_t Z^*_s dM_s- \int^T_t dN_s \] be a solution of the BSDE, let for \(\beta>0\), \[ \widehat H^2_{T,\beta}(\mathbb{R}^d)= \Biggl\{\varphi:\Omega\times \mathbb{R}^+\to\mathbb{R}^d;\text{ predictable},\;|\varphi|^2_\beta= E\int^T_0\exp(\beta A_t)|\varphi_t|^2 dC_t<+ \infty\Biggr\}, \] where \(A_t= \int^t_0\alpha^2_s dC_s\); then there exists a unique solution such that \(|(Y,Z,N)|_\beta< \infty\) for a \(\beta\) large enough. For the entire collection see [Zbl 0866.00052].