Pengalokasian Dan Penjadwalan Salesman Pada Perusahaan Distribusi Dengan Menggunakan Constrained K-Means Clustering

Ramadhanu, Arya (2026) Pengalokasian Dan Penjadwalan Salesman Pada Perusahaan Distribusi Dengan Menggunakan Constrained K-Means Clustering. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6032241132-Master Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

PT.XYZ merupakan perusahaan distributor FMCG yang sedang mengalami tantangan mengenai Sales Territory Management terutama pada pembagian pelanggan dan penjadwalan salesman yang dilakukan secara konvensional sehingga menyebabkan beban kunjungan salesman tidak merata, kunjungan tumpang tindih antar salesman dan jarak tempuh tidak efisien. Penelitian ini bertujuan untuk penerapan algoritma Constrained K-Means Clustering sebagai sistem pendukung keputusan dalam pembagian pelanggan dan penjadwalan salesman serta mengevaluasi hasil klusternya. Penerapan Constrained K-Means Clustering dilakukan 2 tahap, tahap pertama untuk membagi pelanggan dan tahap kedua untuk membuat jadwal salesman. Sample pada penelitian ini adalah 2479 pelanggan yang tersebar di beberapa wilayah Surabaya dan sekitarnya. Kualitas kluster antara sebelum dan sesudah menggunakan Constrained K-Means Clustering dibandingkan dengan menggunakan metrik Silhouette Coefficient dan Davies-Bouldin Index (DBI). Selain itu, perhitungan standar deviasi juga digunakan untuk mengukur beban kerja salesman. Hasil dari Constrained K-Means Clustering menunjukkan bahwa kualitas pembagian pelanggan dan penjadwalan salesman meningkat signifikan. Dibuktikan dengan peningkatan nilai Silhouette Coefficient dari 0,174 menjadi 0,356, penurunan nilai DBI dari 22,55 menjadi 0,87, standar deviasi jumlah pelanggan antar salesman turun dari 29,8 menjadi 0.4 dan rata-rata jarak pelanggan ke centroid turun dari 6,3 km menjadi 4,6 km. Secara manajerial, implementasi sistem pendukung keputusan berbasis kluster dapat membagi wilayah distribusi, meratakan beban kerja salesman dan mengurangi jarak tempuh salesman.
==================================================================================================================================
PT. XYZ is an FMCG distribution company currently facing operational challenges in Sales Territory Management, particularly regarding manual customer allocation and salesman scheduling. This conventional practice leads to an uneven daily workload among salesmen, overlapping operational territories, and inefficient travel distances. This study aims to implement a Constrained K-Means Clustering algorithm as a Decision Support System (DSS) to optimize customer allocation and salesman scheduling, as well as to evaluate the performance of the resulting clusters. The implementation of Constrained K-Means Clustering is executed in a two-stage hierarchical process, where the first stage delineates the macro-level commercial territories for each salesman, and the second stage develops the micro-level tactical daily visit schedules. The sample utilized in this research comprises the geographical coordinates of 2,479 customers dispersed across Surabaya and its surrounding areas. Cluster quality before and after the algorithmic intervention is comparatively evaluated using Silhouette Coefficient and Davies-Bouldin Index (DBI) metrics, while standard deviation calculations are employed to measure the level of workload balancing among the salesmen. The experimental results demonstrate that the application of the Constrained K-Means Clustering algorithm significantly improves the quality of territory allocation and scheduling. This improvement is validated by an increase in the Silhouette Coefficient from 0.174 to 0.356, a sharp decline in the DBI value from 22.55 to 0.87, a reduction in the standard deviation of customer quantities per salesman from 29.8 to 0.4, and a decrease in the average customer distance to the cluster centroid from 6.3 km to 4.6 km. From a managerial perspective, the implementation of this data-driven cluster-based decision support system successfully structures distribution territories, establishes an equitable workload balancing for field personnel, and minimizes salesmen's travel distances.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	FMCG, Sales Territory Management, Constrained K-Means Clustering, Silhouette Coefficient, Davies-Bouldin Index.
Subjects:	T Technology > T Technology (General) > T57.84 Heuristic algorithms.
Divisions:	Interdisciplinary School of Management and Technology (SIMT) > 61101-Master of Technology Management (MMT)
Depositing User:	Arya Ramadhanu
Date Deposited:	27 Jul 2026 02:51
Last Modified:	27 Jul 2026 02:51
URI:	http://repository.its.ac.id/id/eprint/137824

Actions (login required)

View Item