Bug in SystemChain.add_site_dissipation()

[gefux]

Adding a Markovian Lindblad dissipator to a system chain seems to go wrong, as one can see by the decay of the norm in the following example with dissipation on the first of a two site chain (without any coherent evolution):

import oqupy
import numpy as np
import matplotlib.pyplot as plt

sigma_z = oqupy.operators.sigma("z")
sigma_minus = oqupy.operators.sigma("-")
up_dm = oqupy.operators.spin_dm("z+")
down_dm = oqupy.operators.spin_dm("z-")

initial_augmented_mps = oqupy.AugmentedMPS([up_dm, down_dm])
system_chain = oqupy.SystemChain(hilbert_space_dimensions=[2,2])
system_chain.add_site_dissipation(0, sigma_minus, gamma=0.2)

pt_tebd_params = oqupy.PtTebdParameters(
    dt=0.1,
    order=2,
    epsrel=1.0e-6)

pt_tebd = oqupy.PtTebd(
    initial_augmented_mps=initial_augmented_mps,
    system_chain=system_chain,
    process_tensors=[None, None, None, None, None],
    parameters=pt_tebd_params,
    dynamics_sites=[0, 1],
    chain_control=None)

num_steps = 20
results = pt_tebd.compute(num_steps, progress_type="bar")

plt.plot(results['norm'].real)

The norm drops below 0.7 in 20 time steps, which seems to be a real bug and not just a numerical error.