LeMans:

Multi Agent Network Simulations

Artificial-life based simulations are often built upon the concept of agents (or cellular automata) located on a grid of a specified shape (square, infinite, toroidal, 3-dimensional, etc). Interactions are based upon the concept of proximity, defined by the agent neighborhood on the grid. This approach has a number of drawbacks, including dependence of results of a simulation upon arbitrary choices of shape of the world and concept of proximity.
A Multi-Agent Network Simulation is built upon the concept of agents as nodes of several graphs representing the relationships between agents, organizational structures, communication networks, resource dependencies, as well as physical proximity (with the choice of networks dependent on the simulation domain).