A Hybrid Quantum-assisted Column Generation Algorithm for the Fleet Conversion Problem

TL;DR

This paper introduces a hybrid quantum-classical algorithm for the Fleet Conversion problem, leveraging recent quantum techniques to efficiently solve large-scale MWIS problems within a column generation framework.

Contribution

It presents a novel hybrid quantum-assisted approach that combines quantum and classical solvers to address industrial-sized fleet conversion challenges.

Findings

Successfully solved up to 64 tours using quantum-classical methods.

Demonstrated the effectiveness of QUBO representation with fewer qubits.

Showcased potential for quantum algorithms in real-world logistics optimization.

Abstract

The problem of Fleet Conversion aims to reduce the carbon emissions and cost of operating a fleet of vehicles for a given set of tours. It can be modelled as a column generation scheme with the Maximum Weighted Independent Set (MWIS) problem as the slave. Quantum variational algorithms have gained significant interest in the past several years. Recently, a method to represent Quadratic Unconstrained Binary Optimization (QUBO) problems using logarithmically fewer qubits was proposed. Here we use this method to solve the MWIS Slaves and demonstrate how quantum and classical solvers can be used together to approach an industrial-sized use-case (up to 64 tours).