Diagonalizing large-scale quantum many-body Hamiltonians using variational quantum circuit and tensor network

Abstract

Exact diagonalization (ED) is an essential tool for exploring quantum many-body physics but is fundamentally limited by the exponentially-scaled computational complexity. Here, we propose tensor network variational diagonalization (TNVD), which encodes the full eigenenergy spectrum of a quantum many-body Hamiltonian into a matrix product state, and encodes the eigenstates as the evolutions of product states using variational quantum circuit (VQC). Thereby, TNVD reduces the computational complexity of diagonalization from exponential to polynomial in system size $N$. Numerical benchmarks up to $N=100$ spins are provided, which far surpass the computational limit of ED. We further consider quantum Ising model in a random field to reveal the underlying reliance between the efficiency of TNVD and entanglement properties of eigenstates. Typical signs, including the distribution of entanglement entropy (EE) versus eigenenergy and the density of state versus EE, are suggested to indicate area law of entanglement entropy or its violation, which are essential to the TNVD efficiency. Our work establishes TNVD as a powerful and scalable diagonalization approach for large-scale quantum many-body Hamiltonians. The incorporation of VQC lays a promising pathway to applying quantum computation to address the volume-law-EE Hamiltonians that lack efficient classical approaches.