Prediksi Kecepatan Bongkar Muat Kapal Menggunakan Metode Adaptive Group Multi-task Elastic Net Berdasarkan Karakteristik Kapal dan Faktor Operasional

Aurasati, Lukihani (2026) Prediksi Kecepatan Bongkar Muat Kapal Menggunakan Metode Adaptive Group Multi-task Elastic Net Berdasarkan Karakteristik Kapal dan Faktor Operasional. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5003221059-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Proses bongkar muat merupakan aktivitas penting di pelabuhan yang memengaruhi kelancaran arus logistik dan efisiensi operasional terminal. Kecepatan bongkar muat kapal dipengaruhi oleh berbagai faktor yang saling berkaitan sehingga berpotensi menimbulkan multikolinearitas pada pemodelan regresi multivariat klasik. Oleh karena itu, penelitian ini bertujuan untuk membangun model prediksi kecepatan bongkar muat kapal secara simultan pada dua variabel respon, yaitu Box Crane per Hour (BCH) dan Box Ship per Hour (BSH), menggunakan pendekatan Adaptive Group Multi-task Elastic Net. Metode ini mampu melakukan seleksi variabel secara simultan antar respon, dan memungkinkan variabel kategorik dipertahankan sebagai satu kesatuan sehingga tidak dieliminasi secara terpisah. Hasil penelitian menunjukkan bahwa variabel total ITV, working hour, total movement, serta variabel kategorik jenis kapal dan jenis dermaga berpengaruh signifikan terhadap BCH dan BSH, sementara lima variabel lainnya berhasil dieliminasi melalui proses penalti adaptif sehingga menghasilkan model yang lebih sederhana dan efisien. Selain itu, model yang dihasilkan memiliki performa prediksi yang sangat baik dengan nilai η_Λ^2 multivariat sebesar 98,61%, serta nilai R² masing-masing sebesar 81,16% untuk BCH dan 91,28% untuk BSH. Hasil penelitian ini menunjukkan bahwa Adaptive Group Multi-task Elastic Net mampu menghasilkan model prediksi yang stabil dan efektif dalam menangani multikolinearitas dibandingkan regresi multivariat klasik. Implikasi penelitian ini dapat digunakan sebagai dasar pengambilan keputusan operasional pelabuhan, khususnya dalam perencanaan kebutuhan peralatan bongkar muat, penentuan alokasi sumber daya, peremajaan container crane pada dermaga dengan produktivitas rendah, serta pelaksanaan program peningkatan kompetensi operator guna mendukung peningkatan efisiensi bongkar muat kapal.
==================================================================================================================================
Loading and unloading operations are essential activities in ports that affect the smooth flow of logistics and terminal operational efficiency. The speed of cargo handling is influenced by various interrelated factors, which may lead to multicollinearity issues in classical multivariate regression modeling. Therefore, this study aims to develop a simultaneous prediction model for ship loading and unloading performance based on two response variables, namely Box Crane per Hour (BCH) and Box Ship per Hour (BSH), using the Adaptive Group Multi-task Elastic Net approach. This method enables simultaneous variable selection across multiple responses while preserving categorical variables as unified groups, preventing their dummy variables from being eliminated separately. The results indicate that total ITV, working hour, total movement, as well as the categorical variables of vessel type and berth type significantly affect both BCH and BSH, while five other variables were eliminated through the adaptive penalization process, resulting in a more parsimonious and efficient model. Furthermore, the proposed model demonstrated excellent predictive performance, with a multivariate η_Λ^2 value of 98.61%, and R² values of 81.16% and 91.28% for BCH and BSH, respectively. These findings suggest that the Adaptive Group Multi-task Elastic Net method is capable of producing a stable and effective prediction model for handling multicollinearity compared with classical multivariate regression. The results of this study can support operational decision-making in ports, particularly in planning cargo-handling equipment requirements, allocating operational resources, renewing container cranes at low-productivity berths, and implementing operator competency enhancement programs to improve loading and unloading efficiency.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Kecepatan bongkar muat kapal, Multikolinearitas, Adaptive Group Multi-task Elastic Net, Cargo handling productivity, Multicollinearity, Adaptive Group Multi-task Elastic Net
Subjects:	H Social Sciences > HA Statistics H Social Sciences > HA Statistics > HA31.3 Regression. Correlation. Logistic regression analysis.
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Lukihani Aurasati
Date Deposited:	30 Jul 2026 06:19
Last Modified:	30 Jul 2026 06:19
URI:	http://repository.its.ac.id/id/eprint/139241

Actions (login required)

View Item