A TIME-VARIANT FUZZY-STOCHASTIC MARKOV MODEL FOR NON-HOMOGENEOUS WORKFORCE SYSTEMS UNDER EPISTEMIC AND ALEATORY UNCERTAINTY

Abstract

Modern workforce systems operate in highly dynamic environments characterized by structural evolution, uncertain transition mechanisms, and incomplete managerial information. Traditional homogeneous stochastic workforce models often fail to capture real organizational complexities arising from time-dependent policies, ambiguous human behavior, and fluctuating operational conditions. This study proposes a Time-Variant Fuzzy–Stochastic Markov Model for analyzing and optimizing non-homogeneous workforce systems under combined epistemic and aleatory uncertainty. The framework integrates stochastic Markov transition dynamics with fuzzy set theory to simultaneously model randomness inherent in workforce mobility and imprecision arising from subjective assessments, policy ambiguity, and incomplete data. The model introduces time-dependent transition probability structures that accommodate evolving organizational policies, promotion rules, recruitment strategies, and attrition patterns. Aleatory uncertainty is represented through probabilistic state transitions, while epistemic uncertainty is incorporated via fuzzy membership functions describing uncertain managerial judgments and workforce performance evaluations. A hybrid fuzzy–stochastic formulation is developed to estimate workforce distributions across grades over multiple planning horizons. Stability conditions, equilibrium behavior, and system adaptability are analytically examined, demonstrating improved predictive capability compared with classical homogeneous Markov approaches. Simulation experiments illustrate how the proposed framework supports strategic manpower planning, risk-aware decision-making, and adaptive workforce control under uncertain environments. Results show enhanced robustness in forecasting staffing levels, minimizing skill imbalance, and maintaining organizational sustainability despite time-varying disruptions. The proposed model provides a unified analytical foundation for modern workforce analytics where uncertainty arises from both randomness and knowledge limitations. This study contributes a scalable mathematical framework applicable to public institutions, healthcare systems, manufacturing organizations, and technology-driven enterprises seeking resilient workforce optimization under complex uncertainty structures.