Improvement Studies of an Effective Interaction for N=Z sd-shell Nuclei by Neural Networks

The nuclear shell model is one of the successful models in theoretical understanding of nuclear structure. If a convenient effective interaction can be found between nucleons, various observables such as energies of nuclear states are accurately predicted by this model. The basic requirements for the shell model calculations are a set of single particle energies and two-body interaction matrix elements (TBME) which construct the residual interaction between nucleons. This latter could be parameterized in different ways. In this study, we have used a different approach to improve existing USD type Hamiltonians for the shell model calculations of N=Z nuclei in the A=16-40 region. After obtaining the SDNN new effective interaction, shell model calculations have been performed for all N=Z nuclei in sd shell. In which, $^{16}{O}$ doubly magic nucleus has been assumed as an inert core and active particles are distributed in the $d_{5/2}$, $s_{1/2}$ and $d_{3/2}$ single particle orbits. The rms deviations from experimental energy values are lower for the newly generated effective interaction than those obtained using the original one for the studied nuclei.