In terms of excitation I think you want a lumped port. Perhaps a wave port would work? Someone correct me there if I'm wrong. As for your radiation box, I believe you want a box that extends at least lambda/4 from the edges of your antenna and then assign PML boundaries. Good luck.
In wire antenna you should use the lumped element port to connect both end-side of the antenna.
Usually you should have at least a lambda/4 space corresponding to the minimum frequency, around the structure. Be aware of enlarging the radiation box which causes the long simulation time!
At last, after simulation you could plot the s11 parameter to know about the resonant frequency of antenna
You can use a lumped port to connect both ends of the loop. However, the most accurate simulation setup is the one that best represents the measurement setup.