Pradanaputra, Alief Randiansyah (2026) Optimasi Rantai Pasok Kedelai Nasional Multi-Periode Menggunakan Adaptive Large Neighborhood Search. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5002221172-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5002221172-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (13MB) | Request a copy |
Abstract
Kedelai merupakan komoditas pangan strategis bagi Indonesia, namun ketergantungan impor yang mencapai lebih dari 85 persen kebutuhan nasional menjadikan alokasi pasokan dari sumber lokal dan impor sebagai persoalan kritis ketahanan pangan. Distribusi volume impor melalui pelabuhan tertentu ke 38 provinsi dengan tingkat kebutuhan yang berbeda-beda memerlukan perencanaan yang mempertimbangkan biaya, kapasitas, dan batas kebijakan secara simultan. Penelitian ini bertujuan mengoptimalkan alokasi pasokan kedelai nasional multi-periode melalui pendekatan metaheuristik yang mampu menangani skala permasalahan praktis. Masalah diformulasikan sebagai model Mixed Integer Linear Programming (MILP) dengan objektif minimasi total biaya yang mencakup produksi lokal, impor, distribusi pelabuhan, transfer antarprovinsi, inventori, dan kekurangan pasokan selama 12 bulan. Untuk menyelesaikannya, penelitian ini mengimplementasikan Adaptive Large Neighborhood Search (ALNS) dengan representasi solusi berbasis matriks tiga dimensi dan tiga konfigurasi, yaitu ALNS, ALNS-TS, dan ALNS-PRS adapted. Evaluasi komparatif pada 50 percobaan independen menunjukkan ALNS-PRS adapted sebagai konfigurasi terpilih dengan rata-rata kekurangan pasokan mendekati nol. Analisis 16 skenario mengungkapkan bahwa sistem paling rentan terhadap penurunan kapasitas sumber impor dan gangguan berat kapasitas pelabuhan, sementara pengetatan batas ketergantungan impor di bawah 90 persen belum tercapai tanpa peningkatan pasokan lokal. Penelitian ini berhasil mengembangkan model optimasi alokasi pasokan berskala nasional, menghasilkan rekomendasi alokasi berbasis skenario, dan menyajikan keluaran melalui Dasbor interaktif.
==============================================================================================================================================================================================
Soybean is a strategic food commodity for Indonesia, yet import dependency exceeding 85 percent of national demand makes the allocation of local and imported supply a critical food-security issue. Distributing imports through ports to 38 provinces with different demand levels requires simultaneous consideration of cost, capacity, and policy bounds. This study aims to optimize multi-period national soybean supply allocation using a metaheuristic approach capable of handling the problem at a practical scale. The problem is formulated as a 12-month Mixed Integer Linear Programming model that minimizes the costs of local production, imports, port distribution, interprovincial transfers, inventory, and shortage. To solve the problem, this study implements Adaptive Large Neighborhood Search with a three-dimensional matrix-based solution representation and three configurations, namely ALNS, ALNS-TS, and ALNS-PRS adapted. Comparative evaluation over 50 independent trials identifies ALNS-PRS adapted as the selected configuration, achieving near-zero mean shortage. Analysis of 16 scenarios reveals that the system is most vulnerable to reductions in import-source capacity and severe port-capacity disruptions, while tightening the import-dependency bound below 90 percent remains unattainable without increased local supply. This study develops a national-scale supply allocation optimization model, produces scenario-based allocation recommendations, and presents the outputs through an interactive dashboard.
Actions (login required)
 |
View Item |