Enhancing Demand Fulfillment in Metal-Mechanical MSEs Using DDMRP and Holt-Winters with ML Residual Correction

Peru's metalworking MSEs face significant demand fulfillment challenges. CREATEM S.A.C., an automotive spare parts manufacturer, achieves only 47.99% OTIF - well below the 80% industry benchmark - resulting in annual losses of S/ 103,145.40. Since MYPEs make up 99.2% of the enterprises in the metalworking sector, which contributes 11.2% of the national manufacturing GDP, it is critical to optimize their management. The issue has been tackled in earlier studies using forecasting and techniques like DDMRP. Regarding the resources accessible to micro and small businesses, these instruments are constrained. To enhance demand fulfilment, the model combines machine learning algorithms with forecasting techniques like Holt-Winters, exponential smoothing, etc. with the DDMRP approach. Over an eight-month period, simulations were used for validation. The primary outcome was a MAPE of less than 20% and an increase in OTIF to 88%. The outcomes show that the concept works well for enhancing the sector's ability to meet demand.