Short-term Forecasting of Spring Production Volume with Small Data Using Double Exponential Smoothing and Winters’ Method

Abstract

Springs are a crucial part of suspension systems for vehicles. One of the largest companies that produce springs is PT Indospring Tbk that is located in Indonesia. In the last two years, sales volume in PT Indospring Tbk has continuously declined. PT Indospring Tbk needs to take action to prevent further production problems due to inconsistent sales volume. The action that can be taken to prevent production problems, either overproduction or below the customer’s demand, is forecasting the production volume of their product. This paper discussed three forecasting methods which are Double Exponential Smoothing, Additive Winters’, and Multiplicative Winters’ method to determine the number of products that need to be manufactured in the future. Data is gathered from the company's annual report and processed using Minitab 18. By using the same value of parameters, the results show that Winters’ Multiplicative method is the most suitable method for the company with the least estimation of errors (MAPE, MAD, and MSD) compared to others.