Abstract:Variational quantum algorithms and, in particular, variants of the varational quantum eigensolver have been proposed to address combinatorial optimization (CO) problems. Using only shallow ansatz circuits, these approaches are deemed suitable for current noisy intermediate-scale quantum hardware. However, the resources required for training shallow variational quantum circuits often scale superpolynomially in problem size. In this study we numerically investigate what this scaling result means in practice for solving CO problems using Max-Cut as a benchmark. For fixed resources, we compare the average performance of training a shallow variational quantum circuit, sampling with replacement, and a greedy algorithm starting from the same initial point as the quantum algorithm. We identify a minimum problem size for which the quantum algorithm can consistently outperform sampling and, for each problem size, characterize the separation between the quantum algorithm and the greedy algorithm. Furthermore, we extend the average case analysis by investigating the correlation between the performance of the algorithms by instance. Our results provide a step towards meaningful benchmarks of variational quantum algorithms for CO problems for a realistic set of resources.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to evaluate the practical performance of variational quantum algorithms (VQAs) in solving combinatorial optimization (CO) problems. Specifically, the paper focuses on the performance of variational quantum algorithms using shallow - parameterized quantum circuits when solving the Max - Cut problem. The main objectives of the research include: 1. **Relationship between resource requirements and problem scale**: Explore how the number of samples required to train shallow - variational quantum circuits changes as the problem scale increases, especially whether this change has super - polynomial complexity. 2. **Comparison of different algorithms**: Through numerical experiments, compare the average performance of three different algorithms when solving the same problem instance: - Training shallow - variational quantum circuits (VQA) - Random sampling with replacement - Greedy algorithm, starting from the same initial point as the quantum algorithm 3. **Identification of the minimum problem scale**: Determine the problem scales at which the quantum algorithm can consistently outperform random sampling, and for each problem scale, analyze the performance difference between the quantum algorithm and the greedy algorithm. 4. **Analysis of instance - related performance**: Expand the average - case analysis and study the performance correlation of different algorithms on specific instances to understand whether the instances that are difficult for one algorithm are also difficult for another algorithm. Through these studies, the paper aims to provide meaningful benchmark tests for the practical application of variational quantum algorithms in combinatorial optimization problems, especially under the current available noisy intermediate - scale quantum (NISQ) hardware conditions. This helps to understand the potential and limitations of variational quantum algorithms in practical applications.

Benchmarking Variational Quantum Algorithms for Combinatorial Optimization in Practice

Benchmarking Adaptative Variational Quantum Algorithms on QUBO Instances

Variational Quantum Algorithms for Combinatorial Optimization

A benchmarking study of quantum algorithms for combinatorial optimization

Benchmarking of Different Optimizers in the Variational Quantum Algorithms for Applications in Quantum Chemistry

Towards Robust Benchmarking of Quantum Optimization Algorithms

Performance Benchmarking of Quantum Algorithms for Hard Combinatorial Optimization Problems: A Comparative Study of non-FTQC Approaches

A Comparative Study on Solving Optimization Problems With Exponentially Fewer Qubits

Quantum combinatorial optimization beyond the variational paradigm: simple schedules for hard problems

Performant near-term quantum combinatorial optimization

Benchmarking the Variational Quantum Eigensolver using different quantum hardware

Using models to improve optimizers for variational quantum algorithms

Optimization Applications as Quantum Performance Benchmarks

Randomized Benchmarking of Local Zeroth-Order Optimizers for Variational Quantum Systems

Benchmarking Optimizers for Qumode State Preparation with Variational Quantum Algorithms

The Quantum Approximate Optimization Algorithm performance with low entanglement and high circuit depth

Sub-universal variational circuits for combinatorial optimization problems

Discretized Quantum Exhaustive Search for Variational Quantum Algorithms

Filtering variational quantum algorithms for combinatorial optimization

Quantum-Enhanced Greedy Combinatorial Optimization Solver

Quantum Algorithm Exploration using Application-Oriented Performance Benchmarks