Development of a Genetic Algorithm for the search of Optical Model parameters

The analysis of elastic scattering cross sections in terms of the Optical Model is subject to a series of well known ambiguities. Diverse assumptions about the initial values or shape of the potentials frequently produce different parameters, leading to different physical interpretations of the observed data. It would be important to have a starting set of “user independent” optical potentials that fit the experimental data to allow the evaluator to consider a large array of possibilities before committing to a particular optical potential. This work presents a Genetic Algorithm (GA) code that simulates natural selection and evolution, allowing a “blind search” of the multiparametric χ2 surface. In this GA, the genes subject to evolution are the parameters of the optical potential. The GA variables, operators and procedures are described, and the GA is applied to two cases in which the elastic scattering cross section is adjusted: one for the 7Li + 27Al system at energies close to the Coulomb barrier where the interaction occurs near the nuclear surface, and another for the 16O + 16O system where the two nuclei deeply interpenetrate each other. Further developments are described.