Library for solving the customer-to-depot assignment in vehicle routing problems

e419

Abstract

Optimization problems related to vehicle routing planning are very common in modern life, where 

not only is essential for the logistics of supply chains, but also the efficient management of resources 

for public transportation services, solid waste collection, and urban goods distribution. A specific 

case is the Multi-Depot Vehicle Routing Problem (MDVRP), which requires splitting the problem 

into two parts: assigning customers to depots and constructing the routes. The optimal assignment of 

customers to depots is a complex problem in itself, considering capacity and access constraints while 

aiming to minimize the cost of the resulting routes. This work presents BHAVRP, an open-source 

library implemented in Java and Python, which offers a set of algorithms for customer-to-depot 

assignment and can be easily integrated into academic and business solutions due to its flexibility. 

Through a set of computational experiments conducted on 33 classical MDVRP instances, the 

performance of the implemented algorithms was evaluated, with the adaptation of UPGMC standing 

out by achieving the best results in terms of solution quality