I am working on developing a Dissipative Particle Dynamics (DPD) model of a colloid in a bulk fluid. I essentially want to ensure that the perturbations of my meshed colloid are completely uncoupled from the duplicated copies of itself that are present as a result of using periodic boundary condition. This includes any hierarchical interactions. For example, if fluid particles are disturbed from interacting with a duplicated version of the colloid, the propagation of these interactions should dissipate out enough such that by the time this propagation reaches the colloid of interest, that it essentially has a negligible effect on the behavior of the colloid. Unfortunately I myself am not certain about what 'negligible' means, but perhaps someone with more experience in the field could offer some advice on that point in particular.
As a starting point, I would assume the following:
$L-D > r_c$,
where $L$ is the box length, $D$ is the diameter of the colloid and $r_c$ is the largest cutoff for the bead-bead interactions forces. This ensures that the colloid will not interact with itself as a direct result of colloid-colloid interactions, but does not ensure the prevention of the hierarchical interactions I described earlier.