Enhancing Restaurant OTIF Using 5S, Standardized Work, DDMRP, and Machine-Learning-Based Demand Forecasting

Jiménez-Cárdenas P.D.; Orue-Flores P.J.; Chávez-Ugaz R.; Quiroz-Flores J.C. · Advances in Transdisciplinary Engineering (2026) · DOI: 10.3233/atde260683

Resumen

Restaurants operate in highly variable environments, where long waiting times, stockouts, and non-conforming orders deteriorate customer satisfaction and profitability. Prior research has explored Lean tools and forecasting models independently, yet few studies integrate 5S, Standardized Work, DDMRP, and Machine-Learning-Based Demand Forecasting to enhance OTIF performance in restaurant operations. This case study addresses a Peruvian restaurant with an initial OTIF of 6.75%, far below the 92% industry benchmark and generating monthly losses exceeding 3,600 USD. The proposed model reorganizes the kitchen with 5S, stabilizes task execution through standardized work, and ensures supply availability using Random-Forest-based demand forecasting coupled with DDMRP buffers. Validation using time studies, Arena simulation, and inventory analysis increased OTIF to 55.72%, raised OTD by 52.67 percentage points, achieved a Fill Rate of 91.78%, and reduced cycle time to 23.8 minutes. The findings contribute evidence for integrated Lean-analytics approaches in service operations and encourage broader applications across food-service SMEs.

Instituciones: Universidad de Lima