The application of evolutionary computation to the analysis of the profiles of elliptical galaxies: a maximum likelihood approach

Genetic programming technique has been found to be suitable in scenarios where the formulation of models is a data driven process. Evolutionary programming provides a way of searching for parameters in a model without being prone to fall in local minima. A review of how these techniques have been applied to the analysis of elliptical galaxies is given. The effectiveness of a maximum likelihood based fitness function is asserted and is applied to the parameter fitting using evolutionary programming. A maximum likelihood based function is found to show consistent and significant improvement over a hit-based fitness function for modeling the profiles of elliptical galaxies. It is asserted that such a function would potentially improve the quality of model produced by symbolic regression using genetic programming.