Sumantri, Nicholas Joe (2026) Optimasi Biaya Distribusi Pangan Pada Komoditas Daging Ayam Di Indonesia Dengan Menggunakan Grey Wolf Optimizer (GWO). Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5002221003-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5002221003-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (36MB) | Request a copy |
Abstract
Daging ayam merupakan komoditas pangan strategis di Indonesia dengan tingkat konsumsi tinggi, namun distribusi antarwilayah masih terkendala kondisi geografis kepulauan, ketimpangan lokasi sentra produksi, dan ketidakpastian permintaan yang menimbulkan disparitas pasokan serta harga antarprovinsi. Penelitian ini memformulasikan optimasi biaya distribusi daging ayam sebagai model Mixed-Integer Nonlinear Programming (MINLP) pada jaringan hub-spoke, dengan fungsi objektif meminimalkan total biaya yang mencakup produksi, pembelian antarprovinsi, distribusi melalui Central hub, penyimpanan, dan penalti shortage. Model bersifat mixed-integer karena memuat variabel biner aktivasi produksi, dan nonlinear karena biaya produksinya mengikuti economies of scale. Ketidakpastian permintaan dimodelkan melalui pendekatan scenario-based stochastic programming dengan empat skenario, yaitu Normal, High Season, Low Demand, dan Krisis. Model diselesaikan menggunakan Grey Wolf Optimizer (GWO) dan Particle Swarm Optimization (PSO), lalu diperluas dengan Hierarchical Cluster-Based Decomposition (HCD) untuk mengatasi ruang pencarian 35 provinsi, 12 bulan, dan 4 skenario. Hasil eksperimen menunjukkan HCD-GWO memberikan performa terbaik dengan total biaya terbaik sebesar Rp 261,1 triliun dan simpangan baku antar-run sebesar Rp 7,72 triliun. Kondisi baseline didefinisikan sebagai kondisi pra-optimasi, yaitu saat tiap provinsi hanya mengandalkan kapasitas produksi. lokal tanpa redistribusi. Dibandingkan baseline tersebut, HCD-GWO menurunkan total shortage nasional dari 2.928.689 ton·bulan menjadi 667.499 ton·bulan, atau tereduksi 77,2%. Nilai shortage dinyatakan dalam satuan ton·bulan, yaitu volume kekurangan pasokan yang dijumlahkan atas seluruh provinsi dan seluruh bulan serta dibobot probabilitas keempat skenario. Hasil ini menunjukkan bahwa kombinasi model MINLP, pendekatan scenario-based stochastic programming, GWO, dan dekomposisi berbasis cluster efektif menghasilkan strategi distribusi daging ayam yang lebih efisien dan stabil.
=======================================================================================================================================
Chicken meat is a strategic food commodity in Indonesia with a high level of consumption, yet interregional distribution remains constrained by archipelagic geography, uneven production centers, and demand uncertainty that create supply and price disparities across provinces. This study formulates the chicken meat distribution cost optimization problem as a Mixed-Integer Nonlinear Programming (MINLP) model on a hub-spoke network, with an objective function that minimizes the total cost covering production, interprovincial procurement, distribution through a Central hub, storage, and a shortage penalty. The model is mixed-integer because it contains binary production-activation variables, and nonlinear because its production cost follows economies of scale. Demand uncertainty is modeled through a scenario-based stochastic programming approach with four scenarios, namely Normal, High Season, Low Demand, and Crisis. The model is solved using the Grey Wolf Optimizer (GWO) and Particle Swarm Optimization (PSO), and then extended with Hierarchical Cluster-Based Decomposition (HCD) to handle the search space of 35 provinces, 12 months, and 4 scenarios. The experimental results show that HCD-GWO achieves the best performance, with the lowest best total cost of IDR 261.1 trillion and an inter-run standard deviation of IDR 7.72 trillion. The baseline is defined as the pre-optimization condition, in which each province relies only on its local production capacity without redistribution. Compared with that baseline, HCD-GWO reduces the total national shortage from 2,928,689 ton-months to 667,499 ton-months, an 77.2% reduction. The shortage is expressed in ton-months, namely the volume of unmet supply summed over all provinces and all months and weighted by the probability of the four scenarios. These results indicate that combining the MINLP model, the scenario-based stochastic programming approach, GWO, and cluster-based decomposition effectively produces a more efficient and stable distribution strategy for chicken meat in Indonesia.
Actions (login required)
 |
View Item |