Dewi, Fitri Fatma (2026) Peramalan Throughput Harian Reefer Container Menggunakan Metode Temporal Convolutional Network (TCN) untuk Evaluasi Kecukupan Plug: Studi Kasus Terminal Petikemas di Jawa Timur. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5003221031-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5003221031-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (9MB) | Request a copy |
Abstract
Perdagangan maritim yang terus berkembang meningkatkan kompleksitas operasional terminal petikemas, terutama dalam penanganan muatan sensitif suhu yang diangkut menggunakan Reefer Container. Layanan Reefer membutuhkan pasokan listrik yang kontinu melalui Reefer plug, sehingga perencanaan kapasitas plug yang bersifat prediktif menjadi penting untuk meminimalkan risiko gangguan rantai dingin. Penelitian ini bertujuan untuk: (1) meramalkan Throughput harian Reefer Container pada arus internasional (ekspor dan impor) serta domestik (Receiving dan Delivery) untuk ukuran 20 ft dan 40 ft; (2) membandingkan kinerja Temporal Convolutional Network (TCN) dengan metode baseline, yaitu ARIMAX/SARIMAX untuk seri non-intermittent dan Teunter–Syntetos–Babai (TSB) untuk seri intermittent; serta (3) memanfaatkan hasil peramalan tersebut dalam evaluasi kecukupan plug menggunakan indikator Occupancy Rate (OR). Data yang digunakan berupa Throughput harian periode Januari–Juli 2025, dengan Januari–Juni sebagai data in-sample dan Juli sebagai data out-of-sample, serta data alokasi plug periode April–Juli 2025. Karakteristik data dianalisis melalui pengujian tren, musiman mingguan, dan klasifikasi intermittency berbasis ADI–CV². Hasil evaluasi out-of-sample menunjukkan bahwa TCN 1-Head memberikan galat yang lebih rendah dibandingkan ARIMAX/SARIMAX pada seri non-intermittent, sedangkan TCN 2-Head menunjukkan performa yang lebih stabil dibandingkan TSB pada seri intermittent hingga lumpy. Peramalan final periode Agustus 2025 kemudian diintegrasikan dengan skema dwell time berbobot untuk menghitung okupansi plug. Hasil evaluasi menunjukkan bahwa kapasitas plug berada pada kondisi operasional CUKUP sepanjang periode analisis, dengan peluang terjadinya kondisi RISK dan SHORTAGE yang sangat kecil. Temuan ini menunjukkan bahwa integrasi peramalan berbasis TCN dan evaluasi OR berpotensi mendukung perencanaan kapasitas plug dan sistem peringatan dini pada operasional terminal petikemas.
====================================================================================================================================
The continuous growth of maritime trade has increased the operational complexity of Container terminals, particularly in handling temperature-sensitive cargo transported using Reefer Containers. Reefer services require a continuous electricity supply through Reefer plugs, making predictive plug capacity planning essential to minimize the risk of cold chain disruptions. This study aims to: (1) forecast daily Reefer Container Throughput for international flows (export and import) and domestic flows (Receiving and Delivery) for 20-ft and 40-ft Containers; (2) compare the performance of the Temporal Convolutional Network (TCN) with baseline methods, namely ARIMAX/SARIMAX for non-intermittent series and the Teunter–Syntetos–Babai (TSB) method for intermittent series; and (3) utilize the forecasting results to evaluate plug adequacy using the Occupancy Rate (OR) indicator. The study uses daily Throughput data from January to July 2025, with January–June as in-sample data and July as out-of-sample data, along with plug allocation data from April to July 2025. Data characteristics were examined through trend analysis, weekly seasonality testing, and intermittency classification based on the ADI–CV² framework. Out-of-sample evaluation results indicate that the TCN 1-Head model yields lower forecasting errors than ARIMAX/SARIMAX for non-intermittent series, while the TCN 2-Head model demonstrates more stable performance than the TSB method for intermittent to lumpy series. The final forecasts for August 2025 were then integrated with a weighted dwell time scheme to estimate plug occupancy. The evaluation results show that plug capacity remains in a CUKUP (adequate) operational condition throughout the analysis period, with a very low probability of RISK and SHORTAGE conditions. These findings suggest that integrating TCN-based forecasting with OR evaluation has the potential to support plug capacity planning and early warning systems in Container terminal operations.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Reefer Container, Peramalan Intermittent, SARIMAX, TSB, Temporal Convolutional Network, Intermittent Forecast, Reefer Container, SARIMAX, TSB, Temporal Convolutional Network |
| Subjects: | Q Science Q Science > Q Science (General) > Q180.55.M38 Mathematical models Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > QA Mathematics > QA276 Mathematical statistics. Time-series analysis. Failure time data analysis. Survival analysis (Biometry) Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) |
| Divisions: | Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis |
| Depositing User: | Fitri Fatma Dewi |
| Date Deposited: | 02 Feb 2026 07:34 |
| Last Modified: | 02 Feb 2026 07:34 |
| URI: | http://repository.its.ac.id/id/eprint/131629 |
Actions (login required)
 |
View Item |