Algorithm 2025 → present
Algorithm to estimate the adjoint neutron flux via time-reversed neutron random walks
Fixed-source Monte-Carlo problems · with Théophile Bonnet
A new estimator for the adjoint neutron flux based on the many-to-one formula and time-reversed ("backwards") neutron random walks, aimed at faster flux estimation and improved importance sampling / variance reduction for Monte-Carlo transport codes. Validated so far on a toy example; a full paper is in preparation.
- ›Presented as "Time reversions of Neutron Random Walks" at the Sandia · MaThRad workshop (May 2025).
- ›Presented as "Estimating the eigenfunction of the stochastic Neutron Transport Equation" at the 5th SYSORM, Seville.
Asserted Monte CarloVariance reductionIn progress