Adiabatic Quantum Computing with QUBO Formulations
Rui Hua · ResearchSpace (University of Auckland) · 2016
Quantum physics has the potential to revolutionize computer science by building what is called a quantum computer. Quantum computers can perform tasks that are impossible with Turing-machine based classical computers (e.g., generating random sequences of bits) and can outperform classical computers (to different degrees) in a variety of classical computation problems. Several different models of quantum computing exist. Based on the Adiabatic Theorem in physics, Adiabatic Quantum Computing is a novel computing framework that is drastically different from the more traditional quantum logic gate model. Traditionally, quantum computing has only been tested on devices with a very small number of qubits, severely limiting their uses in practice. In recent years however, advances in engineering and physics have allowed the construction for (relatively) large machines called quantum annealers that do what is known as a specific type of adiabatic quantum computation called quantum annealing. Solving classical computation problems with the quantum annealing framework is not an easy task and there are many obstacles to overcome if we want to achieve any kind of speedups over classical algorithms. In this thesis, we present quantum annealing solutions to several classical problems including the Hamiltonian Cycle Problem, Mixed Dominating Set Problem, Graph Isomorphism Problem, Densest k-Subgraph Problem and the Maximum Weighted Independent Set Problem. Experimental results of these quantum solutions on DWave quantum annealers are also reported. We also look at some of the bottlenecks in the quantum annealing framework and propose several novel approach to mitigate their effects on the efficiency of the quantum solutions.