Lecture 12, solutions 8 and 9, exercise 10

Exercises/ex09/Solutions/xxz_groundstate.py

+"""Simple example to find the ground state of the XXZ model with DMRG
+
+.. todo :
+    document...
+"""
+# Copyright 2018 TeNPy Developers
+
+import numpy as np
+import time
+
+from tenpy.models.spins import SpinChain
+from tenpy.networks.mps import MPS
+from tenpy.algorithms.dmrg import run as run_DMRG
+
+
+def run_groundstate_xxz(L=30, Jz=1., hz=0., conserve='best', chi_max=50, Jz_init=None):
+    model_params = dict(
+        L=L, Jx=1., Jy=1., Jz=Jz, hz=hz, bc_MPS='finite', conserve='best', verbose=1)
+    result = {}
+    M = SpinChain(model_params)
+    result['model'] = 'SpinChain'
+    result['model_params'] = model_params
+
+    result['sites'] = np.arange(L)
+    result['bonds'] = np.arange(L - 1) + 0.5
+
+    psi = MPS.from_product_state(M.lat.mps_sites(), (["up", "down"] * L)[:L])
+    dmrg_params = {
+        'trunc_params': {
+            'chi_max': chi_max,
+            'svd_min': 1.e-10,
+            'trunc_cut': None
+        },
+        'mixer': True,
+        'verbose': 1
+    }
+    if Jz_init:
+        # run first time with small hx to break the symmetry
+        model_params_Jz = model_params.copy()
+        model_params_Jz['Jz'] = Jz_init
+        M_Jz = SpinChain(model_params_Jz)
+        dmrg_params['start_env'] = 10,
+        run_DMRG(psi, M_Jz, dmrg_params)
+
+    # run simulation
+    t0 = time.time()
+
+    info = run_DMRG(psi, M, dmrg_params)
+    print("DMRG finished after", time.time() - t0, "s")
+
+    # save results in output file
+    result['chi'] = np.array(psi.chi)
+    result['S'] = np.array(psi.entanglement_entropy())
+    result['E'] = info['E']
+    result['sweeps'] = info['sweep_statistics']['sweep']
+    for key in ['E', 'S', 'max_trunc_err', 'max_E_trunc', 'max_chi']:
+        result_key = 'sweep_' + key
+        result[result_key] = np.array(info['sweep_statistics'][key])
+    return result
+
+
+if __name__ == "__main__":
+    data = run_groundstate_xxz()
+    # here you can do what you need to do with the data - save it, plot it, ...