Optimization of Multi-Product Production Combination Based on Integer Linear Programming in Garment SMEs to Minimize Production Costs

Abstract

Micro, Small, and Medium Enterprises (MSMEs) in the garment sector frequently face production planning challenges arising from demand fluctuations, limited raw materials, capacity constraints, and inadequate inventory management, which ultimately increase operational costs. This study proposes an integrated forecasting–optimization framework combining single exponential smoothing (SES) and integer linear programming (ILP) to determine the optimal multi-product production combination that minimizes total production and inventory costs in a garment MSME in Jiddah. Historical production data of five products over 12 periods were used to forecast demand, which subsequently served as input for the optimization model. The model was solved using LINGO 21 under three scenarios: (1) considering safety stock and capacity constraints, (2) excluding safety stock, and (3) excluding production capacity constraints. The forecasting results produced an average mean absolute percentage error (MAPE) of 5.79%, indicating highly accurate demand estimates. The optimization results showed that Model 2 generated the lowest total cost of IDR 277,755,500 with a production quantity of 8,157 units and no ending inventory. However, Model 1 was identified as the most appropriate scenario for implementation because it maintained a safety stock of 200 units per product while satisfying production and raw material constraints, resulting in a total production of 9,157 units and a total cost of IDR 312,878,500. Furthermore, Model 3 produced the same solution as Model 1, indicating that production capacity was not a binding constraint under the existing operating conditions. Sensitivity analysis confirmed that the safety stock is a critical factor influencing production decisions, inventory levels, and total costs. Theoretically, this study extends conventional ILP applications by integrating demand forecasting and inventory policy considerations into a single decision-making framework for data-limited MSMEs. Managerially, the proposed model provides practical decision support for balancing cost efficiency and inventory security in multi-product production planning.