An Entropy-Based Exploration Strategy in Dynamic PRA

摘要

An entropy-based biasing rule for the guidance of Dynamic PRA simulations is introduced in this paper. The rule aims to continuously adjust itself based on simulation results, in order to guide the simulations towards groups of sequences for which the highest amount of uncertainty exists regarding their end states. The simple rule described in this paper behaves as intended, even though it has limitations that make it unfit for large scale application.