Error with qutip.solver (tolerance too small)

[Niccolò Laurora]

Hi everybody,

I have been trying to simulate the time evolution of the following hamiltonian:

a = tensor(destroy(N), qeye(N))

b = tensor(qeye(N), destroy(N))

H = a*b - a.dag()*b.dag()

for different values of N, such as N = [16, 18, 20, 22, 24, 26]. 
For the time evolution I have been using:

times = np.linspace(0.0, 10.0, 200)

psi0 = tensor(fock(N,0), fock(N,0))

mesolve(H, psi0, times)

However, for N=24 I get the following error:

qutip.solver.integrator.integrator.IntegratorException: Excess accuracy requested. (Tolerances too small.)

Can anyone help me?

Thank you :)

[Éric Giguère]

You can pass options to the solver / integrator:

mesolve(H, psi0, times, options={"method": "adams", "atol": 1e-6, "rtol": 1e-4})

For a full list of options, see:https://qutip.readthedocs.io/en/latest/apidoc/solver.html#api-ode