Implementasi Metode Hybrid Arima dan Anfis untuk Meningakatkan Akurasi Peramalan Supply Tebu dan Perencanaan Supply pada Industri Gula

Putra, Adyatma Taufiq Rahman (2024) Implementasi Metode Hybrid Arima dan Anfis untuk Meningakatkan Akurasi Peramalan Supply Tebu dan Perencanaan Supply pada Industri Gula. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6010202015_Master_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (5MB) | Request a copy

Abstract

Industri gula merupakan salah satu sektor industri vital di Indonesia yang memiliki tantangan besar dalam perencanaan pasokannya. Salah satu tantangan utama adalah besarnya fluktuasi pasokan tebu harian ke pabrik yang dipengaruhi oleh faktor kematangan optimal tebu dan kondisi cuaca. Fluktuasi pasokan tersebut mempengaruhi sulitnya perencanaan penggilingan tebu karena efisiensi mesin dalam pabrik sangat bergantung pada prediksi volume tebu yang akan diolah setiap harinya. Peramalan yang akurat memainkan peran penting karena menjadi dasar input yang krusial bagi perencanaan operasional yang tepat, sehingga meminimalkan ketidakpastian dan meningkatkan efisiensi secara keseluruhan. Oleh karena itu untuk menunjang perencanaan supply tebu pada pabrik gula, penggunaan metode peramalan yang memberikan akurasi yang baik menjadi kunci dalam mengoptimalkan operasional mesin giling sekaligus mempermudah proses perencanaan operasional produksi. Penelitian ini mempertimbangkan dua metode peramalan utama, yaitu Adaptive Neuro-Fuzzy Inference System (ANFIS) dan Autoregressive Integrated Moving Average (ARIMA). Dua metode tersebut dipilih karena pola data pasokan tebu seringkali bersifat l non-linear. ARIMA memiliki fleksibilitas dalam mengikuti pola data historis, namun terbatas dalam menangkap pola non-linear, sedangkan ANFIS mampu menangkap pola-pola ini. Penggabungan kedua metode ini dalam model hybrid diharapkan dapat memberikan hasil yang lebih akurat berdasarkan penelitian sebelumnya. Studi ini menggunakan sebuah studi kasus di PT Kebun Tebu Mas untuk melatih dan menguji model ARIMA, ANFIS dan model Hybrid ARIMA-ANFIS dalam peramalan pasokan tebu. Hasil dari peramalan menggunakan ARIMA dan Hybrid ARIMA-ANFIS, didpatkan hasil MAPE ARIMA sebesar 19,8% dan hasil MAPE Hybrid ARIMA-ANFIS sebesar 12,3%, maka tingkat akurasi metode Hybrid ARIMA-ANFIS lebih akurat daripada ARIMA. Hasil peramalan metode hybrid digunakan untuk perencanaan hasil peramalan yang dihasilkan digunakan sebagai dasar untuk merumuskan perencanaan dan persebaran supply tebu. Fokus utama dari perencanaan ini adalah untuk mengatasi kekurangan supply yang sering terjadi pada awal musim giling, serta mencegah terjadinya oversupply pada puncak musim panen
======================================================================================================
The sugar industry is one of the vital industrial sectors in Indonesia, facing significant challenges in supply planning. One of the main challenges is the large fluctuations in the daily supply of sugarcane to factories, influenced by the optimal maturity of the sugarcane and weather conditions. These supply fluctuations make it difficult to plan sugarcane milling because the efficiency of the machinery in the factories highly depends on the predicted volume of sugarcane to be processed daily. Accurate forecasting plays a crucial role as it serves as a critical input for precise operational planning, thereby minimizing uncertainty and improving overall efficiency. Therefore, to support sugarcane supply planning in sugar factories, the use of forecasting methods that provide high accuracy is key to optimizing the operation of milling machines as well as facilitating the process of operational production planning. This research considers two main forecasting methods, namely the Adaptive Neuro-Fuzzy Inference System (ANFIS) and the Autoregressive Integrated Moving Average (ARIMA). These two methods were chosen because the patterns in sugarcane supply data are often non-linear. ARIMA has the flexibility to follow historical data patterns but is limited in capturing non-linear patterns, whereas ANFIS is capable of capturing these patterns. The combination of these two methods in a hybrid model is expected to yield more accurate results based on previous research. This study uses a case study at PT Kebun Tebu Mas to train and test the ARIMA, ANFIS, and Hybrid ARIMA-ANFIS models in forecasting sugarcane supply. The results of the forecasting using ARIMA and Hybrid ARIMA-ANFIS showed that ARIMA had a MAPE of 19.8% and the Hybrid ARIMA-ANFIS had a MAPE of 12.3%, indicating that the accuracy of the Hybrid ARIMA-ANFIS method is higher than that of ARIMA. The forecasting results from the Hybrid method are used for planning, with the results being used as a basis for formulating the planning and distribution of the sugarcane supply. The main focus of this planning is to address supply shortages that often occur at the beginning of the milling season and to prevent oversupply during the peak harvest season..

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Peramalan, Adaptive Neuro-Fuzzy Inference System (ANFIS), Autoregressive Integrated Moving Average (ARIMA), Pasokan Tebu,Forecasting, Adaptive Neuro-Fuzzy Inference System (ANFIS), Autoregressive Integrated Moving Average (ARIMA), Sugar Supply
Subjects:	Q Science > Q Science (General) > Q180.55.M38 Mathematical models Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. T Technology > T Technology (General) > T174 Technological forecasting T Technology > TS Manufactures > TS176 Manufacturing engineering. Process engineering (Including manufacturing planning, production planning)
Divisions:	Faculty of Industrial Technology > Industrial Engineering > 26101-(S2) Master Thesis
Depositing User:	Adyatma Taufiq Rahman Putra
Date Deposited:	08 Aug 2024 02:23
Last Modified:	08 Aug 2024 02:23
URI:	http://repository.its.ac.id/id/eprint/111306

Actions (login required)

View Item