An Integrated Forecasting and Simulation Framework for Manufacturing Capacity-Risk Assessment

The aim of this study is to examine how data analytics can be used to evaluate manufacturing capacity feasibility under projected demand growth. Daily production data were used to assess output trends, variability, forecasting performance, demand-growth feasibility, and the relationship between labor efficiency and production output. Naïve, ARIMA, and TBATS forecasting models were compared using RMSE, MAE, and MAPE, with TBATS producing the lowest error among the tested models. The selected forecast indicated flat to slightly declining weekly output, below the level required to support a 20% demand increase. To quantify risk beyond the point forecast, the study added a forecast-anchored Monte Carlo simulation that generated plausible future weekly output paths using empirical TBATS residuals and PPHH-based productivity scenarios. Under current productivity, the simulated probability of meeting the 20% demand-growth target was 18.8% per week, and the expected weekly shortfall was 53.70 units. A 15% PPHH improvement raised this probability only to 29.4%, indicating that labor-productivity improvement alone is unlikely to close the capacity-feasibility gap. The findings show that the production system faces both output instability and a high risk of failing to meet projected demand. Managers should prioritize process stabilization and bottleneck diagnosis while evaluating flexible and permanent capacity options using additional cost, downtime, and station-level data.