Reaction Diffusion is a chemical process in which substances react with each other and diffuse at the same time. This can be modeled by a set of partial differential equations to produce the process in a digital landscape. (full design process here)
The diffusion reaction system has several variables that can be manipulated in order to vary the outcome, the first I tried was the diffusion speed. As shown below, the speed of diffusion affects how quickly the chemical reaction occurs. As a side note, for the model I’m using, the reaction must always occur within the bounds of the rectangle.
The next thing I tried to manipulate was the Flow rate. The Flow rate is the rate at which more chemicals are pumped into the system, it also works in tandem with the kill rate, which determines how quickly the chemicals turn into their final state.
Another thing that I have control of in this model is the spawn point for each particle. So until now, all the patterns had been rather symmetrical, but by changing the insertion point, we can completely change the reaction. Alongside doing this step I also changed the amount of iterations the system goes through, hence why you can see a much clearer build up process in this last set of iterations.
While the first four iterations were all below 100, I tried to push the boundaries for the last two, to beyond 600 iterations, and it produced some vastly different results. What I found particularly interesting about this is that because this is a chemical reaction model, I expected the system to reach a state of equilibrium eventually, but even at that point, the system is still evolving.
An unintended consequence of all this was the 3D model of the reaction diffusion system. It clearly demonstrates the build up process which the system goes through. What it appears the simulation does is export the individual iterations as layers, and then creates a solid mesh out of them as the final result.
I think the next step from here will be to try and figure out how the reaction diffusion system can be modeled in 3 dimensions, rather than simply turning the 2D sequence into a 3D model. This I hope will assist me in the steps ahead where an architecture needs to be derived from this physical process.