Forecasting and Production Planning of Packed Oil for Minimizing the Total Cost

This paper aims to forecast weekly demands of four packed oil
products which were 0001A, 0002A, 0003A and 0004B by time series models. Then the model with the least forecast error was applied in
production planning to minimize the total costs, including production cost,
holding cost, and backlog cost. This research devised five
forecast methods, consisting of Moving Average, Simple Exponential Smoothing,
Croston’s, Syntetos and Boylan, and Holt Winter method. The results showed that
Holt Winter method provided the minimum forecast error for product 0001A and
0004B. For product 0002A and 0003A, Croston’s method gave the minimum forecast
error. Then
the forecast results for four weeks in advance were input into a mathematical
program to minimize the total costs. The program was under production capacity
and storage area constraints. The solution obtained was the production plan for
four weeks with the minimal total costs of each product.