Unfolding an under-determined neutron spectrum using genetic algorithm based Monte Carlo

Abstract

An approach of neutron spectrum unfolding based on a stochastic evolutionary search mechanism - Genetic Algorithm is presented. A pulse height spectrum at zero degree from the nuclear reaction of incident energetic carbon ions on a thick silver target carried out at an accelerator facility is used for the unfolding. The problem is also converted to an under-determined one and is unfolded by the present approach and the output is compared with results from the standard FERDOR code. Necessity of carrying out smoothening of the obtained solution is also studied, which leads to approximation of the solution yielding an appropriate solution finally. A simple smoothening and a Monte Carlo averaging, based smoothening methods have been studied. Unfolded flux value obtained after inclusion of the smoothening criteria results in a close agreement to the output obtained from the FERDOR code.