It sounds very similar to SPH ( https://en.wikipedia.org/wiki/Smoothed-particle_hydrodynamic s), but I do not understand the role of the cells. Questions:
- Is the number of particles "almost" the same in all simulation?
- How is every solving step/iteration?. In the "explicit" case:
--- First, are computed the forces that act over every particle?
--- Second, is the movement of every particle computed depending on the forces?

If the problem is similar to this, you can:
- use a fixed list of particles.
- do a terrific parallelization