Design and Development of Sporadic Demands in Prediction for Automobile Industry

The inventory management of spare parts depends heavily on accurate demand forecasting, which is particularly challenging for spare parts due to their intermittent nature. Intermittent demand typically shows up at random and disappears for extended periods of time. For predicting intermittent inventory demand, Croston’s technique is commonly applied. This research introduces three forecasting approaches: the Croston method, its bisection variant, and a new technique. A comparative analysis is conducted on the cumulative means of Croston’s method. Two categories of spare components are chosen for investigation. We adapt Croston’s method to estimate the cumulative mean of demands within a designated lead time. It is demonstrated that the mean square errors satisfy the intermittent demand's theoretical and practical requirements. In light of these the most accurate statistics summary is achievable. The novel technique has demonstrated superior performance in the out-of-sample comparison. Additionally, the findings underscore the reliability of the mean square error as an accurate metric for assessing sporadic demand. Methods: Croston’s method was applied to predict irregular demands. The Cumulative Mean and Bisection Numerical methods were contrasted with the established Croston’s method. The evaluation of results is based on Mean Square Error (MSE) values. Findings: The Cumulative Mean and Bisection Method exhibited MSE values for the Proposed Method that are lower than those of the Existing Method. Novelty: The Cumulative Mean of Demands Method may serve as a more effective approach for forecasting production in the spare parts industries.