Aplikasi Pemilihan Rute Pengiriman Barang pada Perusahaan Elektronik di Surabaya dengan Menggunakan Metode K-Means Clustering Dan Google Maps API

Abstract

Company X is an electronics retailer company operated in Surabaya city. The company is trying to improve their customer satisfaction by doing some evaluations and enhancement on their existing business processes. One of them is the process of goods delivery services. Some shortcomings occur in the process. For example, it is difficult to determine the order of the delivery routes, especially if there are some customers who have preference delivery hours. Another thing, the company cannot provide any information about the time estimation of the goods arrival that makes their customer feel bad about the delivery services. Because of these problems, the company needs a system that is capable of providing the route sequence recommendation and also giving information about estimation of goods arrival.

The system is implemented on website by using codeigniter framework and SQL Server as the database. In this system, the first process is matching the addresses with the response from Google Maps API. This Google Maps API responses are being used in the next process which is calculating route. In order to perform the calculations, Google Maps API has already provided a free of charge service that is capable of giving the best waypoint through the maximum of 8 points addresses. However, the amount of the addresses in one shipment is usually more than 8 addresses. Therefore a method of clustering using K-Means is needed in order to make a group of delivery area into 8 clusters. By then, the route calculation can be done by using the clusters.

The end result of this program is a shipping information system for administrators and drivers of the company. This system is able to provide recommendations for the selection of goods delivery routes and information on the estimated time of goods delivery. Based on some testings, the system can perform route calculations with an average time calculation of 133 seconds for 30 destination routes.