We assume that distribution of electrons follows Fermi-Dirac distribution / statistics in semiconductor model which will help to find the concentrations of electron and holes and the relationship between them. Why is it that we use this distribution? Why in particular is this distribution used in such models?