Integrating Demand Forecasting, Aggregate Planning, and Sensitivity Analysis for Cost-Efficient Production: A Case Study in Furniture Manufacturing

Abstract

Production planning under demand uncertainty remains a critical challenge in make-to-stock manufacturing systems, particularly when forecasting results are not explicitly linked to cost-based planning decisions. This study develops an integrated framework that combines demand forecasting, aggregate planning, and sensitivity analysis to identify a cost-efficient production policy in a furniture manufacturing company. A quantitative case study was conducted using 12 months of historical demand data for S4S products from July 2024 to June 2025. Three forecasting methods, namely Single Exponential Smoothing, Linear Regression, and Holt’s Trend Method, were evaluated using MAPE, MAD, and RMSE. The best-performing forecast was then used as input for aggregate planning under Level and Chase strategies. To assess the robustness of the planning decision, a one-way sensitivity analysis was conducted by varying key cost parameters by ±20%. The results show that Holt’s Trend Method with α = 0.4 and β = 0.1 provided the best overall forecasting performance, with a MAPE of 1.63%, MAD of 67.34 units, and RMSE of 106.08 units. Using this forecast as the demand input, the Chase Strategy generated the lowest total production cost of Rp.185,900,000, compared with Rp.189,523,750 under the Level Strategy. Sensitivity analysis confirmed that the Chase Strategy remained the preferred strategy under all tested cost-parameter scenarios. These findings demonstrate that integrating forecasting validation, aggregate planning, and sensitivity analysis can improve medium-term production planning decisions and provide practical guidance for manufacturing firms facing fluctuating demand.