A New Scaling Function for QAOA Tensor Network Simulations

Abstract

With the rapid development of quantum computers in recent years, the importance of performance evaluation in quantum algorithms has been increasing. One method that has gained attention for performing this evaluation on classical computers is tensor networks. Tensor networks not only reduce the computational cost required for simulations by using approximations but are also deeply connected to entanglement. Entanglement is one of the most important elements for the quantum advantages of quantum algorithms, but the direct relationship between quantum advantages and entanglement remains largely unexplored. Tensor networks are promising as a means to address this question. In this study, we focus on the entanglement in the Quantum Approximate Optimization Algorithm (QAOA). This study aims to investigate entanglement in QAOA by examining the relationship between the approximation rates of tensor networks and the performance of QAOA. Specifically, we actually perform tensor network simulations of QAOA on a classical computer and extend the study of the scaling relations presented in previous research. We have discovered that scaling relations hold even when entanglement entropy is used as the vertical axis. Furthermore, by analyzing the results of the numerical calculations, we propose a new function for the scaling relation. Additionally, we discovered interesting relationships regarding the behavior of entanglement in QAOA during our analysis. This research is expected to provide new insights into the theoretical foundation of the scaling relations presented in previous studies.