Particle Swarm Optimization Algorithm for Vehicle Routing Optimization

Technology has been one of the critical factors behind the industrial revolution. Companies must now
use technological assistance and data processing to produce faster and more efficient business processes. Our case
study is using HDPE Plastic Company, which is in Surabaya, is trying to handle the increasing frequency of shipments.
Due to the rising frequency of shipments, the company is often overwhelmed in handling its loads because no system
can quickly determine the shipping route.

Moreover, other route-determining factors, such as shipment weight, truck capacity, and unique delivery hour requests,
manually add to the routes complexity. The system will run the K-Means cluster function from the database to cluster
all customers in the company. This cluster is one of the factors determining the fitness value in the Particle Swarm
Optimization (PSO) algorithm. After the order data is obtained, the system will use the PSO algorithm to determine the
delivery agenda for each truck. The determining factors of PSO include customer location, priority hours of customer
requests, order weight, and loading capacity of different types of trucks. After obtaining the delivery table of each
truck, the system will use the help of Google Directions Service to determine the routing order from each truck.

The result of this system is a delivery route optimization system that can provide route selection recommendations for
each truck in the company. The system is also able to sort shipments with various shipping priority restrictions. From
the test results, the PSO algorithm in the system can produce routes with less total distance traveled and less travel
duration than the routes generated manually by the employees in the company.