On Numerical Solution to the Problem of Criticality

The contribution deals with numerical solution to the problem of criticality for neutron transport equation using the source iteration method. Especially question of convergence of iterations is examined. It is inferred that a slow convergence can be expected if optical thickness of the space region considered is large. For such a case it is next shown that the results may strongly be affected by the fact that individual iterations are not computed with sufficient accuracy and consequently that the choice of initial source distribution need not influence quality of final result substantially. These findings are illustrated by several results of computations with Monte Carlo code MCNP. Finally another problems are pointed out, which concern application of computing codes based on source iteration method, to criticality analysis.