Pendekatan Berbasis Skenario untuk Pengelolaan Persediaan Menggunakan Simulasi Monte Carlo

Abstract

Importers and distributors of swimming pool chemicals face significant inventory challenges due to highly volatile and unpredictable demand. This study develops a responsive inventory management strategy by integrating multiple regression analysis with Monte Carlo simulation. Monthly sales data (January-December 2024) for three main products (TCCA Powder, Granular, and Tablet) were analyzed. Descriptive statistics revealed extreme demand volatility, with standard deviations exceeding means for all products. While regression models identified significant influences of previous demand, seasonality, and promotions (R² = 0.58-0.78), their forecasting accuracy was poor, as indicated by high Mean Absolute Percentage Error values (59.40%-147.86%). This limitation justified the shift to a probabilistic approach. Monte Carlo simulation using empirical distributions generated wide demand ranges (e.g., 250-22,000 kg for TCCA Granular), enabling the development of scenario-based inventory policies. The study concludes that the integrated regression-simulation framework provides a more realistic foundation for inventory decision-making than deterministic methods alone, particularly under high uncertainty. The primary contribution lies in positioning Monte Carlo simulation primarily as a decision-support tool rather than a forecasting technique, offering companies a practical method to establish dynamic safety stock levels and improve market responsiveness