Gounaris et al. presented a computational method that can be used for the quick screening of zeolite structures and provide predictions regarding which of them have the potential to exhibit high selectivity among a set of molecules of interest. This article builds upon this earlier work and furthers our understanding of diffusion processes in zeolites and other microporous metal oxides. We first present an augmented formulation to account for the flexibility of the zeolitic portal and conduct an analysis to assess the effect of varying the parameters of the associated quadratic potential. We then introduce a methodology to map the energetic landscape, identify all locally optimal conformations, and probabilistically account for the multiplicity of conformers. Finally, we conduct sensitivity analysis on the effective size of the aperture, and show how the methodology can be fine-tuned through experimental observations. A comprehensive database of 290 molecules of industrial interest and a total of 123 different zeolite structures were used in this study. (C) 2009 American Institute of Chemical Engineers AIChE J, 56: 611-632, 2010