mesolve give slightly deferent final state for different tlists

[神屯]

Hi,

I am using mesolve to simulate a time-dependent open quantum system to get the final state of evolution, and calculate its fidelity. By passing different tlists to mesolve and QobjEvo, I get different results:

(1)

tlist=np.linspace(0, T, 2)

fidelity=0.9014736

(2)

tlist=np.linspace(0, T, 8)

fidelity=0.9014740

(3)

tlist=np.linspace(0, T, 10)

fidelity= 0.9079450

(4)

tlist=np.linspace(0, T, 10**5)

fidelity= 0.9079356

It seems like as numbers of points in tlist reach some threshold, the results converge, and if the numbers of points get too small, the ODE integration error arises. I am wondering the reason of this phenomenon, and how I can know if a number of points in tlist is large enough.

Thanks you!