Evaluating importance maps for TRIPOLI-4$^R$ using deterministic or on-line methods
Abstract
However, the estimation of the response of a particle detector in a strongly attenuated radiation field is very computationally
intensive. Radiation-protection problems, by definition, fall into this class, and are therefore intractable with Monte-Carlo
methods unless suitable variance-reduction techniques are applied.
A common need for variance-reduction methods is to evaluate particle tracks with respect to their expected contribution to
the sought detector response. Any function providing such an estimate for each point of the phase space is generally known
as an $importance\ map$. For a certain class of variance-reduction algorithms, the provably optimal choice for the importance
function is represented by the adjoint flux, i.e. the solution of the adjoint fixed-source Boltzmann equation.
.
Origin : Files produced by the author(s)
Loading...