Estimasi Arrival Time Logistik Berbasis Data Automatic Identification System Menggunakan Pendekatan Model Hybrid Long Short-Term Memory – Extreme Gradient Boosting

Daserra, Serafim (2024) Estimasi Arrival Time Logistik Berbasis Data Automatic Identification System Menggunakan Pendekatan Model Hybrid Long Short-Term Memory – Extreme Gradient Boosting. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 2043201021-Undergraduate_Thesis.pdf] Text
2043201021-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2026.

Download (8MB) | Request a copy

Abstract

Transportasi maritim dengan menggunakan kontainer merupakan moda transportasi utama dalam mengangkut barang di Indonesia dibandingkan moda transportasi darat dan udara. Meskipun memiliki banyak keunggulan seperti kapasitas besar, kemampuan menempuh jarak jauh, biaya yang rendah, dan ramah lingkungan, terdapat beberapa permasalahan dalam rantai pasoknya, terutama terkait ketidakpastian waktu kedatangan kapal angkut barang di pelabuhan laut yang berdampak pada perencanaan aktivitas di terminal kapal dan transportasi selanjutnya. Untuk memprediksi waktu kedatangan sebuah kapal di pelabuhan, informasi tentang rute kapal sangat diperlukan. Namun permasalahan lainnya adalah kapal tidak selalu memberikan informasi rute ke pelabuhan. Data yang digunakan dalam memprediksi estimasi waktu kedatangan atau Estimated Time Arrival (ETA) kapal adalah data Automatic Identification System (AIS) kapal, yang berisi informasi tentang posisi, kecepatan, dan arah kapal. Pemodelan untuk memprediksi ETA kapal menggunakan metode machine learning akan dilakukan dengan menggunakan data AIS kapal. Penelitian ini menggunakan studi kasus data AIS di Selat Sunda dan menggunakan teknik hybrid machine learning yaitu LSTM-XGB (Long Short-Term Memory – Extreme Gradient Boosting) yang menghasilkan model prediksi ETA kapal kontainer di Selat Sunda. Tujuan penelitian ini adalah untuk meningkatkan keakuratan prediksi ETA berbasis AIS kapal menggunakan pendekatan machine learning dalam perjalanan ke pelabuhan dengan memanfaatkan rute tujuan yang ditentukan. Dari hasil penelitian ini dapat disimpulkan bahwa model LSTM-XGBoost dengan menggunakan variabel cuaca memiliki performa terbaik dalam memprediksi waktu kedatangan kapal dibandingkan dengan model LSTM dan XGBoost.
===================================================================================================================================
Maritime transportation using containers is the main mode of transportation in transporting goods in Indonesia compared to land and air transportation modes. Although it has many advantages such as large capacity, the ability to travel long distances, low costs, and environmental friendliness, there are several problems in its supply chain, especially related to the uncertainty of the arrival time of freight carriers at sea ports which have an impact on planning activities at ship terminals and subsequent transportation. To predict the time of arrival of a ship in port, information about the route of the ship is indispensable. But another problem is that ships do not always provide route information to port. The data used in predicting the ship's estimated time of arrival (ETA) is the ship's Automatic Identification System (AIS) data, which contains information about the position, speed, and direction of the ship. Modeling to predict ship ETA using machine learning methods will be done using ship AIS data. This research uses a case study of AIS data in the Sunda Strait and uses a hybrid machine learning technique , namely LSTM-XGB (Long Short-Term Memory – Extreme Gradient Boosting) which produces ETA prediction models for container ships in the Sunda Strait. The purpose of this study was to improve the accuracy of ship AIS-based ETA predictions using a machine learning approach on the way to port by utilizing a defined destination route. From the results of this study, it can be concluded that the LSTM-XGBoost model using weather variables has the best performance in predicting ship arrival time compared to LSTM and XGBoost models.

Item Type: Thesis (Other)
Uncontrolled Keywords: ETA, Industri Maritim, Machine Learning, Prediksi, Transportasi, Maritime Industry, Prediction, Transportation
Subjects: H Social Sciences > HE Transportation and Communications
H Social Sciences > HE Transportation and Communications > HE564.A1 Shipping
H Social Sciences > HE Transportation and Communications > HE566.F7 Freighters. Cargo ships
H Social Sciences > HE Transportation and Communications > HE566.T3 Tankers
Q Science > Q Science (General)
Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Q Science > Q Science (General) > Q325.78 Back propagation
Q Science > QA Mathematics
Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
T Technology > T Technology (General) > T58.8 Productivity. Efficiency
Divisions: Faculty of Vocational > 49501-Business Statistics
Depositing User: Serafim Daserra
Date Deposited: 09 Aug 2024 02:48
Last Modified: 04 Sep 2024 04:21
URI: http://repository.its.ac.id/id/eprint/110602

Actions (login required)

View Item View Item