Taqwa, Muhammad Insan (2024) Forecasting Demand for Wheat Flour at PT. XYZ using Time Series Method. Masters thesis, Institut Teknologi Sepuluh Nopember.
![]() |
Text
6032211038-Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (2MB) | Request a copy |
Abstract
Wheat flour is a one of food ingredient which consumed by the public, especially in the noodle and bread industries, it makes wheat flour become important food commodity. Product quality and availability are the main factors in customer satisfaction. The quality of wheat flour is depend on the blending of each type of wheat, while product availability is depend on the amount of stock in the warehouse. Excess inventory has effects on decrease quality due to fumigation process. On the other side, lack of inventory will cause lack of product availability when customers need it, then it can cause lost sales. High fluctuations in wheat flour demand require a forecasting method with a good level of accuracy. In this research, time series forecasting using the Autoregressive Integrated Moving Average (ARIMA) method will be used to predict the need for 5 types of products using monthly historical data for the last 6 years (January 2018 – December 2023). The level of accuracy will use RMSE (Root Mean Square Error), while raw material calculation will use linear regression method. Based on the ARIMA modeling results, the best model for Premium Bread Flour 12.5-13%, using the ARIMA model (2,1,0) with RMSE 164.83, for Premium Noodle Flour 12-12.5% using the ARIMA model (1,1,1 ) with RMSE 176.94, for Premium General Purpose Flour 11-12% using ARIMA (3,1,0) with RSME 11.1, and for Economic Low Protein Flour 9-10% using ARIMA (0,1,1) with RSME 1365.57. For calculating raw materials using linear regression, the correlation between Black Sea Wheat 10% and Economic Low Protein Flour 9-10%, we get R-sq value 91.45%, for Black Sea Wheat 12% and Economic General Purpose Flour 10-10.5 % we get R-sq value 65.17%, for Black Sea Wheat 12% and Premium General Purpose Flour 11-12% we get R-sq value of 46.38%, for Australian Premium Wheat 10.5% and Premium Noodle Flour 12.5 -13%, we get R-sq value 84.92%, for Australian Premium Wheat 10.5% and Premium Bread Flour 13-13.5%, we get R-sq value 51.88%, for Canadian Red Spring Wheat 13.5% and Premium Noodle Flour 12.5-13% we get R-sq value 71.07%, for Canadian Red Spring Wheat 13.5% and Premium Bread Flour 13-13.5% we get R-sq value of 68.50%.
=================================================================================================================================
Tepung terigu merupakan salah bahan baku makanan yang banyak dikonsumsi masyarakat terutama pada industri mie dan roti, sehingga menjadi komoditi pangan yang penting. Kualitas dan ketersediaan produk menjadi faktor utama pada kepuasan pelanggan. Kualitas tepung terigu sangat dipengaruhi oleh blending dari tiap jenis gandum, sedangkan ketersediaan produk ditentukan oleh jumlah stok di gudang. Kelebihan stok menimbulkan efek samping berupa penurunan kualitas akibat fumigasi. Pada sisi lain jumlah stok yang kurang akan berakibat pada kurangnya ketersediaan produk saat diperlukan customer yang mengakibatkan terjadinya lost sales. Fluktuasi permintaan yang cukup tinggi pada tepung terigu membuat diperlukannnya metode peramalan dengan tingkat akurasi yang baik. Pada penelitian ini akan dilakukan peramalan time series dengan metode Autoregressive Integrated Moving Average (ARIMA) untuk meramalkan kebutuhan 5 jenis produk dengan menggunakan data historis bulanan selama 6 tahun terakhir (Januari 2018 – Desember 2023). Tingkat akurasi akan menggunakan RMSE (Root Mean Square Error), sedangkan perhitungan bahan baku akan menggunakan metode regresi linear. Berdasarkan hasil pemodelan ARIMA, didapatkan model terbaik untuk Premium Bread Flour 12.5-13%, menggunakan model ARIMA (2,1,0) dengan RMSE 164,83, untuk Premium Noodle Flour 12-12.5% menggunakan model ARIMA (1,1,1) dengan RMSE 176,94, untuk Premium General Purpose Flour 11- 12% menggunakan ARIMA (3,1,0) dengan RSME 11,1, dan untuk Economic Low Protein Flour 9-10% menggunakan ARIMA (0,1,1) dengan RSME 1365,57. Pada hasil perhitungan bahan baku dengan regresi linear, untuk korelasi antara Black Sea Wheat 10% dengan Economic Low Protein Flour 9-10% didapatkan nilai R-sq 91,45%, untuk Black Sea Wheat 12% dengan Economic General Purpose Flour 10-10.5% didapatlan nilai R-sq 65,17%, Black Sea Wheat 12% dengan Premium General Purpose Flour 11-12% didapatlan nilai R-sq 46,38%, untuk Australian Premium Wheat 10,5% dengan Premium Noodle Flour 12,5-13% didapatlan nilai R-sq 84,92%, untuk Australian Premium Wheat 10,5% dengan Premium Bread Flour 13-13,5% didapatkan nilai R-sq 51,88%, untuk Canadian Red Spring Wheat 13.5% dengan Premium Noodle Flour 12,5-13% didapatlan nilai R-sq 71,07%, untuk Canadian Red Spring Wheat 13.5% dengan Premium Bread Flour 13-13,5% didapatkan nilai R-sq 68,50%.
Item Type: | Thesis (Masters) |
---|---|
Uncontrolled Keywords: | ARIMA, Forecasting, Linear Regression Time Series, Wheat Flour |
Subjects: | Q Science Q Science > Q Science (General) Q Science > Q Science (General) > Q180.55.M38 Mathematical models Q Science > QA Mathematics |
Divisions: | Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT) |
Depositing User: | Muhammad Insan Taqwa |
Date Deposited: | 13 Aug 2024 08:45 |
Last Modified: | 14 Aug 2024 03:17 |
URI: | http://repository.its.ac.id/id/eprint/115382 |
Actions (login required)
![]() |
View Item |