Genetic-Algorithm-based Light Curve Optimization Applied to Observations of the W UMa star BH Cas

Abstract

I have developed a procedure utilizing a Genetic-Algorithm-based optimization scheme to fit the observed light curves of an eclipsing binary star with a model produced by the Wilson-Devinney code. The principal advantages of this approach are the global search capability and the objectivity of the final result. Although this method can be more efficient than some other comparably global search techniques, the computational requirements of the code are still considerable. I have applied this fitting procedure to my observations of the W UMa type eclipsing binary BH Cassiopeiae. An analysis of V-band CCD data obtained in 1994/95 from Steward Observatory and U- and B-band photoelectric data obtained in 1996 from McDonald Observatory provided three complete light curves to constrain the fit. In addition, radial velocity curves obtained in 1997 from McDonald Observatory provided a direct measurement of the system mass ratio to restrict the search. The results of the GA-based fit are in excellent agreement with the final orbital solution obtained with the standard differential corrections procedure in the Wilson-Devinney code.