Evaluating Quantum Cire Cutting for QAOA: Performance Benchmarks in Ideal and Noisy Environments

At present, quantum computers operate in the Noisy Intermediate-Scale Quantum (NISQ) era facing significant limitations in qubit count and error rates. Quantum circuit cutting techniques offer a promising solution by decomposing quantum circuits into smaller sub-circuits that can be executed on current hardware. In this study, the effectiveness of wire cutting is investigated both under ideal and noisy conditions. First, several wire-cutting techniques were compared using a Greenberger-Hornze-Zeilinger (GHZ) circuit. Randomized measurements wire cut consistently achieved higher accuracy and scaled more favorably with increased shot budgets compared to Pauli-based wire cutting. Next, the randomized measurements wire cut method was applied to Quantum Approximate Optimization Algorithm (QAOA) circuits solving the Max- Cut problem. The performance of wire-cut QAOA against uncut QAOA in both ideal and noisy simulation environments were compared. PennyLane was used for ideal simulations and Qiskit Aer with the IBM Brisbane noise model for noisy conditions. Wire cutting enabled the execution of larger QAOA circuits with fewer qubits, but reduced solution accuracy compared to uncut circuits. Under ideal circumstances, accuracy loss can be mitigated by increasing the shot budget. However, under noisy conditions, the benefits of increased shot budgets are less consistent and are primarily observed for problems with a single wire-cut. For larger problems, the impact of increased shots under noisy simulations remains inconclusive. In conclusion, the randomized measurements wire cut method is a promising method to reduce the number of required qubits, but it has its limitations under noisy conditions