A genetic algorithm approach to fitting interferometric data of post-AGB objects: I. the case of the Ant nebula

Abstract

We present GADRAD, a Python module that adopts heuristic search techniques in the form of genetic algorithms, to efficiently model post-asymptotic giant branch (post- AGB) disc environments. GADRAD systematically constructs the multi-dimensional pa- parameter probability density functions that arise from the fitting of radiative transfer and geometric models to optical interferometric data products. The result provides unbiased descriptions of the object's potential morphology, component luminosities and temperatures, dust composition, disc density profiles and mass. Correlation in the estimated parameters as well as potential degeneracies are revealed. Estimated probability distributions of the post-AGB environment parameters provide insight into the shaping processes that may occur in the transition from the post-AGB to the planetary nebula phase. We test parameter recovery on simulated artificial data products of a typical post-AGB environment. We then use GADRAD to model the mid-infrared spectrum and visibilities of the Ant nebula (Mz3), taken with the Very Large Telescope Interferometer's instrument MIDI. Our result is consistent with a large dusty disc with similar parameter values to those previously found by Chesneau et al., except for a larger dust mass of $3.5^{+7.5}_{-2.2}\times10^{-5}$ M$_{\odot}$. The parameter confidence intervals determined by GADRAD, can however be relied upon to impose additional constraints on all disc and system parameters. Based on our analysis and other considerations, we tentatively suggest that Mz3 is a pre-PN ejected during a magnetic (polar) common envelope interaction, where the binary may or may not have survived at the core of the nebula.