Implementation of the reverse/adjoint Monte Carlo method into Geant4

We have implemented the reverse/adjoint Monte Carlo method into Geant4. In this method the primary particles are tracked backward from the sensitive part of the geometry till the external source surface by following the reverse processes. By this way the computing time is spent only for tracks that are contributing to the tallies and the reverse Monte Carlo method is much more rapid than the normal forward Monte Carlo method for simulation cases where the sensitive part is small compared to the rest of the geometry and compared to the extended external source. In this paper we first present the theoretical principles of the Reverse Monte Carlo method. Then we describe how this method and the reverse processes have been implemented into Geant4. Finally we compare and discuss the simulation results obtained with the reverse and forward Monte Carlo modes in Geant4.