Traditional bioequivalence (BE) assessment relies primarily on Non-Compartmental Analysis (NCA), which cannot differentiate physiological variability from product variability. Building upon an emerging framework that integrates Physiologically Based Biopharmaceutics Modeling (PBBM) with population estimation methods, this project evaluates the potential of combining PBBM with the Stochastic Approximation Expectation Maximization (SAEM) algorithm to explore alternatives in BE assessment. Utilizing desvenlafaxine as a case study, the project aims to deconvolute participant-specific physiological variability from actual formulation-specific variability to provide an enhanced analysis of drug product performance. The methodology comprised three distinct stages: developing and validating a whole-body PBPK model for desvenlafaxine, integrating this model with the saemix algorithm to establish a population framework; and executing model-based bioequivalence (MBBE) simulations under various variability scenarios. The PBPK model successfully described the pharmacokinetics of desvenlafaxine. Findings further demonstrate that the integration of PBBM and SAEM is feasible for population-level estimations, though refining structural strategies for inter-occasion variability (IOV) remains necessary to fully optimize the approach. MBBE simulations demonstrated that combined physiological and formulation-specific variability increased the predicted Type II error from 0.7% to 4.7% for an inherently bioequivalent product. Conversely, for a non-bioequivalent drug product, combined variability decreased the predicted Type I error to 4.7%, compared to 6.0% under isolated formulation-specific variability. Ultimately, this work provides a preliminary evaluation of model-informed methodologies that aim to offer deeper mechanistic insights into variability origins, potentially supporting more robust regulatory decisions for mass-produced drug products following further refinement.