This simulation demonstrate's HASH's message passing ability between agents.
Here, a grid of agents is moving randomly and remarking on their neighbors. Whenever an agent has a neighbor, it sends that neighbor a message with its own color information.
Similarly, an agent will listen to all of the colors of neighbors that it hears from, and adapt its own color to be the average of itself and its reported colors.
As the simulation runs, we'll see the agents converge on a common average color, but with distinct patches where agents have been separated by physical space.
By modifying the search_radius specified for agents in the initial state, you can modify the rate at which convergence occurs.