Number-Projected Coherent-State Kernels for Variational Bose-Field Dynamics

Reproducibility package for the paper "Number-projected coherent-state kernels for variational Bose-field dynamics" submitted to New Journal of Physics (2026). Contains the LaTeX manuscript source, Python benchmark scripts, numerical data (CSV), and publication-quality figures. Scripts reproduce all results (Figs. 1–5) for two-, three-, and four-mode Bose-Hubbard models using number-projected coherent states with fixed-basis Galerkin and moving-basis TDVP methods. Requirements: Python 3.8+, NumPy, SciPy, Matplotlib.本复现包对应2026年提交至《新物理学杂志》（New Journal of Physics）的论文《数投影相干态核用于变分玻色场动力学》（Number-projected coherent-state kernels for variational Bose-field dynamics）。该包包含LaTeX稿件源文件、Python基准测试脚本、CSV格式数值数据，以及符合出版标准的高质量插图。本脚本可复现二模、三模及四模玻色-哈伯德（Bose-Hubbard）模型的全部结果（图1至图5），所用方案结合数投影相干态与固定基伽辽金（Galerkin）法、移动基时域变分原理（Time-Dependent Variational Principle，TDVP）法。运行依赖：Python 3.8及以上版本，以及NumPy、SciPy、Matplotlib库。