Splitting Schemes for Backward Stochastic Differential Equations

Abstract

This paper concerns splitting methods for solving backward stochastic differential equations (BSDEs). By splitting the original $d$-dimensional BSDE into $d$ BSDEs and approximating these split BSDEs, we propose splitting schemes for the BSDE. The splitting schemes are rigorously analyzed and first-order error estimates are theoretically obtained. Numerical tests are given to verify the theoretical results.