Pengembangan Model Optimasi Rute Distribusi Berbasis Vehicle Routing Problem Pada Transportation Management System Pt XYZ

Abstract

Distribusi produk pangan beku memerlukan perencanaan rute yang efisien untuk menjaga ketepatan waktu pengiriman, pemanfaatan armada, dan keberlangsungan cold chain. Proses penyusunan rute secara manual berpotensi menghasilkan jarak tempuh yang lebih panjang dan penggunaan sumber daya yang kurang optimal. Penelitian ini bertujuan mengembangkan model optimasi rute berbasis Capacitated Vehicle Routing Problem with Time Windows (CVRPTW) yang diimplementasikan pada backend Transportation Management System (TMS) menggunakan Google OR-Tools. Model mempertimbangkan kapasitas kendaraan, jendela waktu pelayanan, jam operasional, serta zonasi geografis sebagai soft constraint, dengan matriks jarak dan waktu yang dibangun menggunakan Open Source Routing Machine (OSRM). Sistem diuji melalui pengujian fungsional, validitas model, dan integrasi sistem, kemudian dievaluasi menggunakan data operasional perusahaan dengan membandingkan hasil optimasi terhadap penjadwalan manual. Hasil penelitian menunjukkan bahwa sistem menghasilkan solusi yang memenuhi seluruh kendala operasional, menurunkan total jarak tempuh sebesar 31,1%, mengurangi total waktu operasional sebesar 26,6%, serta mempertahankan pemenuhan jendela waktu pelayanan. Hasil tersebut menunjukkan bahwa sistem layak digunakan sebagai pendukung pengambilan keputusan distribusi yang lebih efisien dan konsisten.

Frozen food distribution requires efficient route planning to ensure timely deliveries, optimal fleet utilization, and cold chain continuity. Manual route planning often results in longer travel distances and inefficient resource utilization. This study aimed to develop a Capacitated Vehicle Routing Problem with Time Windows (CVRPTW) optimization model implemented within the backend of a Transportation Management System (TMS) using Google OR-Tools. The proposed model considers vehicle capacity, customer time windows, operating hours, and geographical zoning as a soft constraint, while travel distance and time matrices are generated using the Open Source Routing Machine (OSRM). The developed system was evaluated through functional testing, model validation, and system integration testing, followed by a performance evaluation using the company's operational distribution data by comparing the optimized routes with the existing manual planning process. The results demonstrate that the proposed system satisfies all operational constraints, reduces total travel distance by 31.1%, decreases total operational time by 26.6%, and maintains compliance with customer service time windows. These findings indicate that the developed system is suitable for supporting more efficient, consistent, and data-driven distribution planning.