Benchmark

Weighted Max-Cut Benchmark

QAOA / Quantum Walk Optimisation Algorithm (QWOA) · Optimization · 31 qubits · Qiskit, Cirq

Extended Max-Cut benchmark using weighted graph instances, which present a harder optimization landscape than the standard unweighted variant. Weighted instances introduce a proliferation of poor local optima and exacerbate barren-plateau issues. Recent work benchmarks the non-variational Quantum Walk Optimisation Algorithm against two classical local-search heuristics on weighted instances up to 31 nodes, reporting more favourable scaling with problem size.

Max-Cutweighted-graphsQAOAcombinatorial-optimizationQWOA

3 credible sources · last verified 6 months ago

Extended Max-Cut benchmark using weighted graph instances, which present a harder optimization landscape than the standard unweighted variant. Weighted instances introduce a proliferation of poor local optima and exacerbate barren-plateau issues. Recent work benchmarks the non-variational Quantum Walk Optimisation Algorithm against two classical local-search heuristics on weighted instances up to 31 nodes, reporting more favourable scaling with problem size.

Key Metrics
Max graph size
31 nodes
Methods compared
QAOA vs QWOA (quantum walk)
Why It Matters

Demonstrates that weighted instances are substantially harder than unweighted MaxCut, with barren plateaus and local optima proliferating in the optimization landscape.

Hardware

Simulator / hardware-agnostic

Framework

Qiskit, Cirq