← / methods
// 05
Monte Carlo vs Analytical Closure
Why stochastic sampling is not the only path — and why analytical closure changes the deployment economics of advanced reactors.
Monte Carlo criticality codes (MCNP, Serpent, OpenMC) converge at O(1/√N) in the number of histories. Reaching pcm-level precision on a full core burnup requires hours to days on hundreds of cores.
An analytical spectral closure reaches the same precision in a single dense solve on a workstation. That collapses the design-iteration loop from weeks to minutes.
The practical consequence is regulatory: licensing packages can carry deterministic error bounds instead of statistical confidence bands.