Hybrid variational quantum eigensolvers: merging computational models

Variational quantum eigensolvers (VQEs) are a highly successful technique for simulating physical models on quantum computers. Recently, they were extended to the measurement-based approach of quantum computing, bringing the strengths and advantages of this computational model to VQEs. In this work, we push the design and integration frontiers of VQE further by blending measurement-based elements into the gate-based paradigm to form a hybrid VQE. This facilitates the design of a problem-informed variational ansatz and also allows the efficient implementation of many-body Hamiltonians on NISQ devices. We experimentally demonstrate our approach on a superconducting quantum computer by investigating the perturbed planar code, Z2 and SU(3) lattice gauge theories, and the LiH molecule.