I've played with the Gray Scott Model a lot over the last year, but I've never been bold enough to actually tinker with the equations themselves.
Until now.
I've often wondered how a reaction diffusion system would behave with the introduction of some trigonometric terms into the model. So after some tinkering with different parts of the equations, I've ended up changing the reaction rate part of the equation which was:
uv2
To use sine and cosine:
cos(u) * sin(v) * v
The u and v values are clipped between zero and one complete turn in radians,
I've also changed the standard five point Laplacian stencil to a skewed nine point stencil.
After some playing with the new model, there are some apparent differences. It seems that having the trigonometric component introduces some banding effects, although other phenomena such as spiral wavefronts can still be created:
The demo application lives here and the source code is available here.
I've also changed the standard five point Laplacian stencil to a skewed nine point stencil.
After some playing with the new model, there are some apparent differences. It seems that having the trigonometric component introduces some banding effects, although other phenomena such as spiral wavefronts can still be created:
The demo application lives here and the source code is available here.