Majority Voting with Recursive QAOA and Cost-restricted Uniform Sampling for Maximum-Likelihood Detection in Massive MIMO

The data repository contains data sets obtained with Quantum-approximate optimization algorithm (QAOA) simulations and experiments for the main article in [1]. For a comprehensive understanding, please check readme.pdf file and refer to the main article in [1]. We design, theoretically model and simulate QAOA-MVSIC algorithm combining QAOA, majority voting (MV) and successive interference cancellation (SIC) to target experimental challenges of QAOA for maximum-likelihood (ML) decoding in n × n massive multi-input multi-output (MIMO) systems. QAOA experiments are performed in IBM Eagle r3 quantum processor with 127 qubits in IBM Qiskit Runtime (QR). Simulation results for quantum circuits are obtained by using Qiskit. We simulate Binary Phase Shift Keying (BPSK) and Quadrature Phase Shift Keying (QPSK) modulated massive MIMO decoding with QAOA-MVSIC for n = 24 while experiments are performed for n in the interval [17, 64]. QAOA circuit output strings for simulations and experimental tests on the IBM Eagle r3 processor, along with measured values including cost based parameters are provided for all instances. Simulation and experimental parameters, including optimized angles, error mitigation settings, circuit diagrams and problem-specific values including channel gain matrix and transmitted symbol are provided.[1] B. Gulbahar, “Majority Voting with Recursive QAOA and Cost-restricted Uniform Sampling for Maximum-Likelihood Detection in Massive MIMO”, TecxhRiv, 2024.