Abstract:We explore strategies aimed at reducing the amount of computation, both quantum and classical, required to run the Quantum Approximate Optimization Algorithm (QAOA). First, following Wurtz et al. [<a class="link-external link-http" href="http://Phys.Rev" rel="external noopener nofollow">this http URL</a> A 104:052419], we consider the standard QAOA with instance-independent "tree" parameters chosen in advance. These tree parameters are chosen to optimize the MaxCut expectation for large girth graphs. We provide extensive numerical evidence supporting the performance guarantee for tree parameters conjectured in [<a class="link-external link-http" href="http://Phys.Rev" rel="external noopener nofollow">this http URL</a> A 103:042612] and see that the approximation ratios obtained with tree parameters are typically well beyond the conjectured lower bounds, often comparable to performing a full optimization. This suggests that in practice, the QAOA can achieve near-optimal performance without the need for parameter optimization. Next, we modify the warm-start QAOA of Tate et al. [Quantum 7:1121]. The starting state for the QAOA is now an optimized product state associated with a solution of the Goemans-Williamson (GW) algorithm. Surprisingly, the tree parameters continue to perform well for the warm-start QAOA. We find that for random 3-regular graphs with hundreds of vertices, the expected cut obtained by the warm-start QAOA at depth $p \gtrsim 3$ is comparable to that of the standard GW algorithm. Our numerics on random instances do not provide general performance guarantees but do provide substantial evidence that there exists a regime of instance sizes in which the QAOA finds good solutions at low depth without the need for parameter optimization. For each instance studied, we classically compute the expected size of the QAOA distribution of cuts; producing the actual cuts requires running on a quantum computer.

Theoretical Approximation Ratios for Warm-Started QAOA on 3-Regular Max-Cut Instances at Depth $p=1$

Warm-Started QAOA with Custom Mixers Provably Converges and Computationally Beats Goemans-Williamson's Max-Cut at Low Circuit Depths

Warm-Started QAOA with Aligned Mixers Converges Slowly Near the Poles of the Bloch Sphere

The Better Solution Probability Metric: Optimizing QAOA to Outperform its Warm-Start Solution

Bridging Classical and Quantum with SDP initialized warm-starts for QAOA

Systematic study on the dependence of the warm-start quantum approximate optimization algorithm on approximate solutions

Lower Bounds on Number of QAOA Rounds Required for Guaranteed Approximation Ratios

Strategies for running the QAOA at hundreds of qubits

An angle rounding parameter initialization technique for ma-QAOA

A hybrid quantum-classical approach to warm-starting optimization

Quantum Approximate Optimization Algorithms for Maxmimum Cut on Low-Girth Graphs

Local classical MAX-CUT algorithm outperforms $p=2$ QAOA on high-girth regular graphs

A Recursive Lower Bound on the Energy Improvement of the Quantum Approximate Optimization Algorithm

The QAOA gets stuck starting from a good classical string

Tight Lieb-Robinson Bound for approximation ratio in Quantum Annealing

Recursive greedy initialization of the quantum approximate optimization algorithm with guaranteed improvement

Recursive QAOA outperforms the original QAOA for the MAX-CUT problem on complete graphs

QAOA with fewer qubits: a coupling framework to solve larger-scale Max-Cut problem

Quantum approximate optimization algorithm with random and subgraph phase operators

Quantum annealing initialization of the quantum approximate optimization algorithm

Tight Lieb–Robinson Bound for approximation ratio in quantum annealing