Pemodelan Optimasi Rute Armada Pengangkut Daur Ulang Sampah Plastik Kota Surabaya Menggunakan K-Means Clustering dan Multiple-Depot Vehicle Routing Problem (MDVRP)

Aryasatya, Muhammad Tsar Gibran (2024) Pemodelan Optimasi Rute Armada Pengangkut Daur Ulang Sampah Plastik Kota Surabaya Menggunakan K-Means Clustering dan Multiple-Depot Vehicle Routing Problem (MDVRP). Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 5010201161-Undergraduate_Thesis.pdf] Text
5010201161-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2026.

Download (3MB) | Request a copy

Abstract

Daur ulang merupakan salah satu aktivitas manajemen sampah plastik dengan cara memilah, membersihkan, menghancurkan, dan membentuknya kembali ke produk jadi yang bisa digunakan. Tantangan utama yang dihadapi Dinas Lingkungan Hidup Kota Surabaya dan pengepul sampah plastik PET Kota Surabaya saat ini adalah tidak ada pembagian pemukiman dan penerapan rute optimal yang jelas untuk pengambilan sampah daur ulang oleh collection center pemerintah dan swasta. Sistem pengambilan sampah plastik melibatkan armada kendaraan yang berkeliling di pemukiman untuk mengambil sampah plastik lalu mengatarkannya ke collection center/depot/pengepul. Untuk menyelesaikan permasalahan yang kompleks dengan total titik customer (BSU) Kota Surabaya adalah sejumlah 355 titik, diperlukan pendekatan cluster first, route second untuk memperoleh rekomendasi rute optimal dengan waktu komputasi yang terjangkau. Penelitian ini menggunakan metode K-Means Clustering untuk mengelompokkan 355 bank sampah unit berdasarkan lokasi geografisnya secara akurat. Selanjutnya, setiap klaster diselesaikan dengan model Vehicle Rounting Problem (VRP) atau Multiple Depot Vehicle Routing Problem (MDVRP) untuk mencari rute paling optimal dari segi biaya transportasi. Implementasi K-Means Clustering menghasilkan klasterisasi dengan jumlah klaster terbaik adalah lima klaster berdasarkan metrik performa Davies-Bouldin Index, distortion score, dan silhouette coefficient. Model optimasi MDVRP linear programming menghasilkan rekomendasi rute independen untuk setiap klaster dengan hasil biaya transportasi yang ditanggung pengepul dapat dihemat 53% hingga 63% di semua klaster dengan utilisasi kendaraan >84%. Hasil klastering dan optimasi MDVRP (termasuk batasan dan asumsi yang digunakan) juga telah divalidasi, diverifikasi, dan dianalisis secara profitabilitas yang menunjukkan net present value positif dengan payback period < 5 tahun, sehingga rekomendasi rute yang diberikan dapat diterapkan di sistem nyata dan tidak merugikan antar-pengepul.
==============================================================================================================================
Recycling is a crucial activity in plastic waste management that involves sorting, cleaning, crushing, and reforming plastic into usable finished products. The main challenge currently faced by Surabaya City’s Dinas Lingkungan Hidup (DLH) and waste collectors is the absence of clear optimal routes for the collection of recyclable waste by both government and private collection centers. The plastic waste collection system involves a fleet of vehicles that travel through residential areas to collect plastic waste and deliver it to collection centers or depots. Addressing the complex issue of managing Surabaya City's 355 collection points requires a cluster first, route second approach to achieve optimal routing recommendations with reasonable computational time. This study employs the K-Means clustering method to accurately group 355 waste collection units based on their geographic locations. Each cluster is then optimized using the Vehicle Routing Problem (VRP) or Multiple Depot Vehicle Routing Problem (MDVRP) models to determine the most cost-effective transportation routes. The implementation of K-Means clustering identified five optimal clusters based on the Davies-Bouldin Index, distortion score, and silhouette coefficient performance metrics. The MDVRP MILP optimization model provided independent routing recommendations for each cluster, resulting in collection centers’s transportation cost savings of 53% to 63% across all clusters with the vehicle utilization is >84%. The clustering results and MDVRP optimization (including the constraints and assumptions used) have been validated, verified, and analyzed in terms of its profitability (the result is positive net present value while the payback period is <5 year), ensuring that the recommended routes can be applied in the real system without disadvantaging any collection centers.

Item Type: Thesis (Other)
Uncontrolled Keywords: Sampah plastik PET, vehicle routing problem, multiple depot, rute, K-means clustering, bank sampah, PET plastic waste, route, waste bank
Subjects: H Social Sciences > HE Transportation and Communications > HE336.R68 Route choice
Divisions: Faculty of Industrial Technology and Systems Engineering (INDSYS) > Industrial Engineering > 26201-(S1) Undergraduate Thesis
Depositing User: Aryasatya Muhammad Tsar Gibran
Date Deposited: 19 Aug 2024 05:08
Last Modified: 19 Aug 2024 05:08
URI: http://repository.its.ac.id/id/eprint/110185

Actions (login required)

View Item View Item