MONTE CARLO SHIELDING CALCULATIONS USING EVENT-VALUE PATH-LENGTH BIASING.

The role of the event-value and the point-value functions in importance sampling is clearly defined. A biasing technique that alters the transport kernel using the event-value function is developed. This path-length biasing technique is demonstrated by application to a deep penetration shielding problems. The results indicate a substantial reduction in computational cost over the standard exponential transform or adjoint flux biasing of the collision site selection.