I used autodock and at first when I added distribution to autodock for the first time I saw ligand and protein very close together. So how can the distance or position of the ligands be changed with respect to the protein?
In the process of docking, the relative position and orientation of receptor and ligand are randomly changed to generate a large number of different starting poses, these are then processed to find any poses that show interaction energies indicative of complex formation. Ideally, you'll find that the resulting final poses with the best interaction energies converge towards one specific solution.