Abstract:Quantum computing provides powerful algorithmic tools that have been shown to outperform established classical solvers in specific optimization tasks. A core step in solving optimization problems with known quantum algorithms such as the Quantum Approximate Optimization Algorithm (QAOA) is the problem formulation. While quantum optimization has historically centered around Quadratic Unconstrained Optimization (QUBO) problems, recent studies show, that many combinatorial problems such as the TSP can be solved more efficiently in their native Polynomial Unconstrained Optimization (PUBO) forms. As many optimization problems in practice also contain continuous variables, our contribution investigates the performance of the QAOA in solving continuous optimization problems when using PUBO and QUBO formulations. Our extensive evaluation on suitable benchmark functions, shows that PUBO formulations generally yield better results, while requiring less qubits. As the multi-qubit interactions needed for the PUBO variant have to be decomposed using the hardware gates available, i.e., currently single- and two-qubit gates, the circuit depth of the PUBO approach outscales its QUBO alternative roughly linearly in the order of the objective function. However, incorporating the planned addition of native multi-qubit gates such as the global Molmer-Sorenson gate, our experiments indicate that PUBO outperforms QUBO for higher order continuous optimization problems in general.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is whether the Polynomial Unconstrained Binary Optimization (PUBO) form is superior to the Quadratic Unconstrained Binary Optimization (QUBO) form when using the Quantum Approximate Optimization Algorithm (QAOA) to solve continuous optimization problems. Specifically, the researchers experimentally compared the performance differences between PUBO and QUBO in solving continuous optimization problems, especially in terms of solution quality, parameter training efficiency, and circuit width and depth. The core issue of the paper is to explore the advantages and applicability of PUBO relative to QUBO under current and future quantum hardware conditions. ### Main Contributions 1. **Implemented a QAOA algorithm capable of solving any polynomial optimization problem**: This implementation allows control of the bit depth used and the domain of the input variables. 2. **Conducted in - depth case studies on two established continuous optimization benchmark functions**: Evaluated the performance of PUBO and QUBO problem forms on these two functions. ### Research Background - **Application of quantum computing in optimization problems**: Quantum computing provides powerful algorithmic tools and has already shown superiority over classical solvers in specific optimization tasks. - **QAOA algorithm**: QAOA is an algorithm based on the quantum gate model that can simulate the process of adiabatic quantum computing and is suitable for solving high - order polynomial optimization problems. - **Differences between PUBO and QUBO**: PUBO can directly handle high - order polynomial problems, while QUBO needs to convert high - order terms into quadratic terms, which usually requires the introduction of auxiliary qubits. ### Experimental Methods 1. **Discretization of the objective function**: Discretize the continuous objective function \( f : [a, b] \to \mathbb{R} \) into a binary representation. 2. **Conversion of the objective function to a Hamiltonian**: Convert the discretized objective function into a Hamiltonian in quantum mechanics. 3. **Implementation of the Hamiltonian using quantum gates**: Construct a quantum circuit to simulate the time evolution of the Hamiltonian. ### Experimental Results - **Solution quality**: PUBO performs significantly better than QUBO on high - order functions. Especially on the 2D - Rosenbrock function, the performance of QUBO tends to saturate as the number of QAOA iterations increases, while PUBO continues to benefit from a higher number of iterations. - **Parameter training**: The two methods perform similarly in terms of the number of parameter training steps, but PUBO has a shorter running time when using a classical quantum circuit simulator because it requires fewer qubits. - **Circuit width and depth**: PUBO has a higher circuit depth on current hardware, but after the implementation of multi - qubit gates in the future, its circuit depth will be comparable to or even lower than that of QUBO. ### Conclusion The experimental results show that for high - order continuous polynomial optimization problems, the PUBO form is superior to the QUBO form in terms of solution quality and parameter training efficiency. However, PUBO has a higher circuit depth on current hardware, which is an important consideration in practical applications. In the future, with the implementation of multi - qubit gates, the advantages of PUBO will be more obvious.

Evidence that PUBO outperforms QUBO when solving continuous optimization problems with the QAOA

Quadratic versus Polynomial Unconstrained Binary Models for Quantum Optimization illustrated on Railway Timetabling

Boosting quantum annealing performance through direct polynomial unconstrained binary optimization

Solving QUBOs with a quantum-amenable branch and bound method

Evidence of Scaling Advantage for the Quantum Approximate Optimization Algorithm on a Classically Intractable Problem

Combinatorial optimization with quantum imaginary time evolution

A Multilevel Approach For Solving Large-Scale QUBO Problems With Noisy Hybrid Quantum Approximate Optimization

Quadratic unconstrained binary optimization problem preprocessing: Theory and empirical analysis

Solving Quadratic Unconstrained Binary Optimization with divide-and-conquer and quantum algorithms

Solving Combinatorial Optimization Problems with a Block Encoding Quantum Optimizer

Towards a universal QAOA protocol: Evidence of a scaling advantage in solving some combinatorial optimization problems

Numerical Evidence for Exponential Speed-up of QAOA over Unstructured Search for Approximate Constrained Optimization

Optimized QUBO formulation methods for quantum computing

PO-QA: A Framework for Portfolio Optimization using Quantum Algorithms

Enhancing Quantum Algorithms for Quadratic Unconstrained Binary Optimization via Integer Programming

Optimum-Preserving QUBO Parameter Compression

Quantum Graph Optimization Algorithm

Solving boolean satisfiability problems with the quantum approximate optimization algorithm

Local to Global: A Distributed Quantum Approximate Optimization Algorithm for Pseudo-Boolean Optimization Problems

How to Approximate any Objective Function via Quadratic Unconstrained Binary Optimization