Pemodelan Dynamic Vehicle Routing Problem untuk Last-Mile Delivery pada Quick Commerce Berbasis Dark Store

Yuwandina, Rayshanda (2026) Pemodelan Dynamic Vehicle Routing Problem untuk Last-Mile Delivery pada Quick Commerce Berbasis Dark Store. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5010221082-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Layanan quick commerce seperti Astro beroperasi menggunakan model dark store dengan komitmen pengiriman 15-30 menit. Praktik manual terdesentralisasi untuk penggabungan pesanan (batching) dan penentuan rute saat ini memicu inefisiensi jarak serta rentan terhadap pelanggaran Service Level Agreement (SLA) saat volume pesanan melonjak. Penelitian ini merancang model keputusan berbasis Dynamic Vehicle Routing Problem (DVRP) guna mengelola pesanan dinamis last-mile delivery secara sistematis. Evaluasi membandingkan tiga model yaitu Model 1 baseline (on demand dispatch), Model 2 (DVRP dengan batching), dan Model 3 (DVRP dengan integrasi batching serta transshipment menggunakan transfer point berbasis K-Means clustering). Eksperimen numerik pada 27 skenario memvariasikan tingkat permintaan, jumlah armada kurir, dan durasi batch window (3, 5, dan 7 menit). Hasil simulasi membuktikan Model 3 menghasilkan performa terbaik dengan rata-rata penghematan jarak 14,04% dan penurunan biaya operasional 20,19% dibandingkan kondisi baseline. Selain itu, jumlah armada kurir menjadi faktor paling signifikan dalam mempertahankan kepatuhan SLA. Oleh karena itu, direkomendasikan penyesuaian armada sesuai tingkat permintaan, standar batch window 5 menit, dan alokasi 3 titik transfer per dark store guna menyeimbangkan efisiensi biaya dan target pencapaian SLA.
===================================================================================================================================
Quick commerce services like Astro operate using a dark store model with 15 to 30 minutes delivery commitments. Current decentralized manual practices for order batching and routing trigger mileage inefficiencies and are vulnerable to Service Level Agreement (SLA) violations when order volumes surge. This study designs a decision model based on the Dynamic Vehicle Routing Problem (DVRP) to systematically manage dynamic last-mile delivery orders. The evaluation compares three models: Model 1 as the baseline (on-demand dispatch), Model 2 (DVRP with batching), and Model 3 (DVRP integrating batching and transshipment using K-Means clustering-based transfer points). Numerical experiments across 27 scenarios varied demand levels, courier fleet sizes, and batch window durations (3, 5, and 7 minutes). Simulation results prove that Model 3 yields the best performance, generating an average mileage savings of 14.04% and a 20.19% reduction in operational costs compared to the baseline condition. Furthermore, the courier fleet size is the most significant factor in maintaining SLA compliance. Therefore, it is recommended to adjust the fleet size according to demand levels, establish a standard 5-minute batch window, and allocate three transfer points per dark store to balance cost efficiency and SLA achievement targets.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Dynamic Vehicle Routing Problem, Quick Commerce, Dark Store, Last-Mile Delivery, Transshipment.
Subjects:	T Technology > T Technology (General) > T57.62 Simulation T Technology > T Technology (General) > T57.83 Dynamic programming T Technology > T Technology (General) > T57.84 Heuristic algorithms. T Technology > T Technology (General) > T58.62 Decision support systems
Divisions:	Faculty of Industrial Technology > Industrial Engineering > 26201-(S1) Undergraduate Thesis
Depositing User:	Rayshanda Yuwandina
Date Deposited:	05 Aug 2026 09:35
Last Modified:	07 Aug 2026 07:28
URI:	http://repository.its.ac.id/id/eprint/144098

Actions (login required)

View Item