|
abstract:
Reflected backward SDEs (RBSDEs) can be considered as generalizations of optimal stopping problems, where the process, which can be exercised, additionally depends on the optimal value process. Inspired by primal and dual bounds for the value of optimal stopping problems, we present a method to construct confidence intervals for the value of a class of discrete RBSDEs. This method requires that an approximate solution for the RBSDE has already been pre-computed by some arbitrary numerical method. In this respect we suggest a Monte-Carlo method which can be considered as a generalization of the Longstaff-Schwartz algorithm to RBSDEs. |
