Evaluating intermittent demand forecasting techniques for spare part supply chains

Effective forecasting of intermittent demand is critical for optimizing spare part inventory management due to its sporadic and erratic nature. This study assesses and compares various forecasting methods tailored for such demand patterns, including Croston’s method, exponential smoothing, ARIMA, neural networks, Facebook Prophet, Syntetos-Boylan Approximation, and moving average models. Historical demand data from a real-world supply chain is analyzed to evaluate accuracy and bias, identifying the most suitable model for practical applications. The findings provide valuable insights for improving inventory planning, minimizing costs, and enhancing supply chain efficiency, contributing to both academic knowledge and practical advancements in spare part management.