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.

QAOA of the Highest Order

Quantum Approximate Optimisation Applied to Graph Similarity

An adaptive quantum approximate optimization algorithm for solving combinatorial problems on a quantum computer

Quantum Approximate Optimization With Parallelizable Gates

Quantum approximate optimization algorithm with random and subgraph phase operators

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

The QAOA with Few Measurements

Phantom Edges in the Problem Hamiltonian: A Method for Increasing Performance and Graph Visibility for QAOA

Performance of the Quantum Approximate Optimization Algorithm on the Maximum Cut Problem

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

Analytical Expressions for the Quantum Approximate Optimization Algorithm and its Variants

Analyzing the quantum approximate optimization algorithm: ansätze, symmetries, and Lie algebras

Short-depth QAOA circuits and quantum annealing on higher-order ising models

Strategies for running the QAOA at hundreds of qubits

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

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

Deep-Circuit QAOA

Efficient Classical Computation of Quantum Mean Values for Shallow QAOA Circuits

Evaluating Quantum Approximate Optimization Algorithm: A Case Study

Quantum Approximate Optimization Algorithm with Cat Qubits