Grid-Based Simulation and the Method of Conditional Least Squares

K. B. Ensor and P. W. Glynn

Proceedings of the 1996 Winter Simulation Conference, 325-331 (1996)

This paper is concerned with the use of simulation to compute the conditional expectations that arise in the method of conditional least squares. Our approach involves performing simulations at each point on a discrete grid imbedded within a statistical parameter space. Our main result concerns the number of grid points and amount of simulation necessary in order to obtain a degree of accuracy comparable to that in the case in which the conditional expectations are available in closed form.