Gunhee Park, Caltech

Efficient embedding framework for tensor network influence functional of Anderson impurity model

We develop a quantum embedding theory for constructing a tensor network-based influence functional method in quantum impurity models, such as the Anderson impurity model. Here, we formulate a tensor network method that can project bath degrees of freedom into the embedding bath orbital space. The original discrete-time formalism is extended to the continuous-time formulation, which is expressed by the differential equation of motion in the embedding space. This framework leads to the algorithm that requires only linear computational costs to the number of timesteps with smaller timestep errors. The numerical performance is compared in the quench dynamics of the single impurity Anderson model.