Ramadhan, Hanif Hisyam (2026) Optimasi Vehicle Routing Problem Untuk Produk Perishable Dengan Customer Satisfaction Berbasis Quality Decay Menggunakan Hybrid Genetic Algorithm Dan 2-Opt Pada Distribusi Ikan Segar. Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 5010221118-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
5010221118-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy |
Abstract
Distribusi ikan kembung (Rastrelliger kanagurta) sebagai produk perishable menghadapi tantangan penurunan kualitas (quality decay) yang bersifat eksponensial akibat paparan waktu dan suhu selama pengiriman, sehingga dibutuhkan model optimasi rute yang mengintegrasikan aspek kualitas produk secara eksplisit. Penelitian ini mengembangkan model MT-CVRPTW-PG (Multi-Trip Capacitated Vehicle Routing Problem with Time Windows for Perishable Goods) melalui sintesis model Multi-Trip CVRP dan model quality decay dari referensi terdahulu, dengan fungsi objektif meminimalkan total biaya distribusi yang mencakup biaya tetap kendaraan, biaya transportasi, dan penalti penurunan kualitas. Model diselesaikan menggunakan algoritma Hybrid GA+2-Opt tanpa operator crossover, dengan konfigurasi parameter optimal P_{2Opt}\ =\ 0.3,\ P_{m}\ =\ 0.5, dan elitism rate = 0.20 yang ditetapkan melalui full factorial design atas 48 kombinasi. Validasi menggunakan solver CBC pada skala 7 pelanggan mengonfirmasi seluruh solusi feasible dengan rata-rata ARPD 8.48% dan algoritma mampu menemukan solusi hingga 167 kali lebih cepat dari CBC pada instance dengan time window sempit. Pengujian dilakukan pada 10 instance Solomon Benchmark dengan empat skala dataset (25 – 100 pelanggan), dibandingkan dengan GA Murni, ACO Murni, dan Hybrid ACO+2-Opt menggunakan empat metrik: kualitas solusi, konsistensi, waktu komputasi, dan robustness. Hybrid GA+2-Opt unggul secara statistik atas GA Murni dan ACO Murni (p = 0.003) dan merupakan algoritma paling andal dengan total valid 10/10 hingga skala 75 pelanggan, cocok untuk instance dengan pola clustered dan random di seluruh skala. Hybrid ACO+2-Opt menghasilkan Best Z setara (selisih rata-rata 2%, p = 0.086) dengan keunggulan spesifik pada pola mixed, namun dengan waktu komputasi rata-rata 960 detik per run di skala 100 pelanggan. Seluruh pelanggan menerima produk dengan nilai kualitas \mathrm{Q}\geq Q_{min}=0.80 di semua instance yang diuji, mengonfirmasi bahwa model quality decay yang dikembangkan efektif menjaga kesegaran produk sepanjang rantai distribusi.
=====================================================================================================================================
Fresh mackerel (Rastrelliger kanagurta) distribution faces the challenge of exponential quality deterioration (quality decay) caused by time and temperature exposure during delivery, necessitating a routing optimization model that explicitly integrates product quality. This study develops an MT-CVRPTW-PG (Multi-Trip Capacitated Vehicle Routing Problem with Time Windows for Perishable Goods) model by synthesizing a Multi-Trip CVRP model and a quality decay model from prior literature, with an objective function minimizing total distribution cost comprising fixed vehicle costs, transportation costs, and quality deterioration penalties. The model is solved using a Hybrid GA+2-Opt algorithm without a crossover operator, with optimal parameters P_{2Opt}\ =\ 0.3,\ P_{m}\ =\ 0.5, and elitism rate = 0.20 determined through full factorial design over 48 combinations. Validation using the CBC solver at a 7-customer scale confirmed all solutions feasible with an average ARPD of 8.48%, and the algorithm found solutions up to 167 times faster than CBC on tight time window instances. Experiments were conducted on 10 Solomon Benchmark instances across four dataset scales (25–100 customers), benchmarked against Pure GA, Pure ACO, and Hybrid ACO+2-Opt using four metrics: solution quality, consistency, computation time, and robustness. Hybrid GA+2-Opt was statistically superior to Pure GA and Pure ACO (p = 0.003) and proved the most reliable algorithm with a perfect valid rate of 10/10 up to 75 customers, performing best on clustered and random instance types across all scales. Hybrid ACO+2-Opt achieved comparable Best Z values (2% average gap, p = 0.086) with specific advantages on mixed-type instances, but at a computational cost averaging 960 seconds per run at the 100-customer scale. All customers received products with quality value Q≥Qmin=0.80 across all tested instances, confirming the effectiveness of the developed quality decay model in preserving product freshness throughout the distribution chain.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | 2-Opt, Hybrid Genetic Algorithm, Kepuasan Pelanggan, Produk Perishable, Quality Decay, Vehicle Routing Problem. 2-Opt, Customer Satisfaction, Hybrid Genetic Algorithm, Perishable Goods, Quality Decay, Vehicle Routing Problem. |
| Subjects: | Q Science > Q Science (General) > Q337.3 Swarm intelligence Q Science > QA Mathematics > QA402.5 Genetic algorithms. Interior-point methods. Q Science > QA Mathematics > QA402.6 Transportation problems (Programming) T Technology > T Technology (General) > T57.6 Operations research--Mathematics. Goal programming T Technology > T Technology (General) > T57.84 Heuristic algorithms. |
| Divisions: | Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis |
| Depositing User: | Hanif Hisyam Ramadhan |
| Date Deposited: | 21 Jul 2026 04:49 |
| Last Modified: | 21 Jul 2026 04:49 |
| URI: | http://repository.its.ac.id/id/eprint/135976 |
Actions (login required)
 |
View Item |