Optimizing Global Polyethylene Production Planning Based on Demand - Supply Using Pearson Correlation and Linear Regression

The primary objective of this study is to optimize global polyethylene (PE) production planning by reducing the risks of over-inventory and shortage-inventory, which arise from persistent mismatches between supply and demand. These imbalances have long created inefficiencies in the PE industry, leading to increases in storage and holding costs and lost sales opportunities. To address this problem, a predictive, data-driven framework was developed by applying industrial engineering tools. A quantitative methodology was employed, Pearson correlation analysis was used to measure the strength of the relationship between global PE demand and production, followed by the construction of a simple linear regression model to forecast production volumes. Model validation was conducted using the coefficient of determination (R²) and Mean Absolute Percentage Error (MAPE). Once validated, the regression model was integrated into Master Production Schedule (MPS) and Material Requirements Planning (MRP) systems to evaluate its operational impact. The results demonstrate a very strong correlation (r = 0.992) and high predictive accuracy (R² = 98.45%, MAPE = 1.33%). Integration into MPS and MRP reduced inventory deviation from 1.47% (comprising 1.23% over-inventory and 0.24% shortage-inventory) to 0.63% (with complete elimination of over-inventory and a manageable shortage level), generating potential inventory cost savings of USD 4.60–7.66 billion. The innovation of this research lies in embedding regression-based forecasting into MPS and MRP subsystems within a production planning system, providing a transparent, interpretable, and replicable framework for production planning. The contribution is twofold: (1) demonstrating problem solving using industrial engineering methods to minimize inventory risks and optimize resource allocation, and (2) establishing a scalable analytical foundation that can be further enhanced through advanced AI systems and hybrid forecasting to address non-linearity, seasonality, and market uncertainty in the global polymer industry.