Is there a way to maintain receptor symmetry when performing flexible ligand docking protocol?
If you're looking to dock the ligand symmetrically (that is, with one ligand per asymmetric unit), you might want to try setting up symmetry prior to doing ligand docking. (I don't think it's been tested, but there's a non-zero chance that it would work.)
If you're attempting to dock a ligand in a way that breaks symmetry, there's not an easy way to otherwise maintain symmetry. We have some restraints which do a soft enforcement of symmetry (that is, penalize but don't prohibit symmetry-related atoms from breaking symmetry.), but that's mainly of residue identities. There's not a general way which I'm aware of which allows you to conformationally restrain symmetric partners without actually invoking the symmetry mechanism.