Logistics is a promising industry that has proliferated over the years. With growing
operation sizes, the challenges in the supply chain have become more complex. Even
minor improvements over the existing delivery routes can help companies significantly
cut expenses. Current classical computers cannot optimally solve today’s real-world
applications, which created a commercial interest in quantum computing technologies
that are potentially more capable than classical methods for these challenges. This
Master’s Thesis studies whether this interest is justified by solving Capacitated Vehicle
Routing Problem (CVRP) with Quantum Approximate Optimization Algorithm (QAOA),
a quantum computing algorithm designed to solve combinatorial optimization problems.
The aim is to optimize the Hamiltonian formulation of CVRP, whose ground state
encodes the optimal solution. This paper finds QAOA with its Ansatz variant in a
column generation setting as a viable solution that scales well with increasing problem
sizes. To this end, our Hamiltonian formulations are classically simulated in a problem
with 12 customers and a single depot, which found near-optimal results. As far as we
are aware, this problem is also currently the biggest CVRP instance that was solved by
QAOA.
«
Logistics is a promising industry that has proliferated over the years. With growing
operation sizes, the challenges in the supply chain have become more complex. Even
minor improvements over the existing delivery routes can help companies significantly
cut expenses. Current classical computers cannot optimally solve today’s real-world
applications, which created a commercial interest in quantum computing technologies
that are potentially more capable than classical methods for these challe...
»