While my mind races with ideas on how to expand and make this a complex-and-interesting-enough simulation, I'm for now trying to come up with a "minimum viable product" and then iterate from there.
I'm basically trying to decide on the cell "types", and how the interaction is done during the simulation. For a start, I'm considering a basic set of : plains, water, mountains and forest.
Unlike Conway's game of life, the simulation wont happen per "cycle", but will be more realtime. I don't want tiles to change suddenly, but to allow the effect of neighboring tiles to affect them gradually. When a certain threshold is met, then the tile changes.
The question now is how to make each tile affect tiles around it, and what are the conditions and triggers to begin affecting tiles.
I've been thinking in a way of how certain tiles help others to thrive (in terms of life), damage it, or are neutral.
For instance, a water tile is beneficial to a forest tile. But perhaps if there is too much water surrounding just one forest tile, it might turn it into something else, like a swamp (forest+water). In this case, it gets really complex, because I need to consider how each tile might affect other tiles individually, not just in terms of being positive or negative.
One way to do this was to give each tile a certain quantity of each of the natural elements, and hardcode how much each type has of each natural element. Let's say: forest has "10% water", "20% earth", "70% wood", while a water tile has "100% water". So the water tile next to a forest tile, would GIVE Water*100*Z(some variable here), and receive Earth*0.2*Y + Wood*0.7*Y
In the meantime, I decided to change the gridmap a bit, and is now a bit more "hex-like", each tile has 6 neighbors instead of 8
