Rigid protein docking deals with the protein / receptor as rigid structure while in induced fit docking receptor sidechains are allowed to flex during refinement of the poses when should i use each one in docking?
Seems like you are awfully familiar with either IFD or rigid docking.
In every case you had better use IFD if your softwares and CPU powers overwhelmingly support it, as the IFD is not as low-time consuming as the normal docking.
It depends what kind of systems (ligand and receptor) you have, although with flexible docking (i..e, IFD) you may mainly get diverse and correctly predicted results; in some cases you may have risk to obtain false positive results. Thus, for every docking poses you need to investigate carefully if crucial residues are in close contacts with your ligand or not.