Integrating DPMO, Machine Learning, Lean Six Sigma, and a simulation-based Digital Twin framework for enhancing operational performance in higher education systems

Higher Education Institutions (HEIs) are progressively challenged to improve both operational efficiency and academic performance in the face of raising demand, resource constraints, and quality expectations. These challenging operational and academic impediments require robust improvement methodologies. This study proposes an integrated framework combining Lean Six Sigma (LSS), Discrete-Event Simulation (DES), Random Forest (RF) Machine Learning (ML), and a conceptual Digital Twin (DT) framework to analyze and improve administrative processes in HEIs. The proposed approach employs DES to model critical academic and administrative processes, identify bottlenecks, and assess the impact of improvement strategies under varying scenarios of the student enrollment process. A key contribution of this research is the incorporation of Defects Per Million Opportunities (DPMO) as a quantitative performance metric within the simulation environment, enabling precise evaluation of key performance indicators in process variability and quality levels. By embedding DPMO calculations into the simulation-based model, the developed approach provides a data-driven mechanism to monitor performance improvements and benchmark process capability. An RF ML surrogate model is employed to approximate the behavior of the validated DES, providing rapid prediction of performance outcomes and quantifying the predictive importance of process variables associated with defect rates and system efficiency. A mixed-methods case study is tailored to a real case of HEI to validate the applicability of the proposed approach, demonstrating significant improvements in process efficiency, service quality, and resource utilization. The results highlight the effectiveness of integrating DPMO-based evaluation within simulation-driven LSS initiatives to support continuous improvement in HEIs. The simulation experiments demonstrate that the proposed framework has the potential to significantly improve process performance, with the best improvement scenario achieving a Sigma Level of 5.22 while virtually eliminating process defects under the evaluated operating conditions. The proposed framework provides a comprehensive decision-support tool that integrates process improvement, predictive analytics, and sustainability considerations for smart and data-driven HEI management.