Empirical Analysis of Diffusive Load-Balancing

In the diffusive model, each vertex is allowed to send an arbitrary amount of its load to its neighbors. If we have discrete load it is hard to distribute the load among these evenly.
We take a look at two different approaches to deal with this problem, vertex-based and edge-based algorithms. Random, quasirandom and deterministic algorithms of both categories are compared in several simulations on Hypercube and Torus graphs.