TETRIS-ADAPT-VQE: An adaptive algorithm that yields shallower, denser circuit ansätze

Recent progress on variational quantum eigensolver (VQE) algorithms has reduced the resource requirements with an adaptive algorithm for problem-tailored ansatz construction called ADAPT-VQE. The algorithm uses local energy gradient information to iteratively construct a VQE ansatz, one unitary at a time. In this work, we introduce a variation of the algorithm dubbed TETRIS-ADAPT-VQE which modifies the way the ansatz is constructed by adding multiple unitaries with disjoint supports in each iteration. Our algorithm results in denser but significantly shallower circuits, without increasing the number of CNOT gates or variational parameters. Moreover, the expensive step of measuring the energy gradient with respect to each candidate unitary is performed only a fraction of the time. These improvements bring us closer to the goal of demonstrating a practical quantum advantage on quantum hardware.