Generation

While creating algorithms for random continent generation, I came across an old algorithm I had worked with called Coupled Map Lattice. Using that to create a continent was hard but I was able to accomplish it doing layers of filters: generate map with noise between -1 and 1 apply CML using Spatiotemporal Intermittency settings (a = 1.75, ε = 0.6) I found that using something too chaotic didn’t have a noticeable affect compared to just generating noise in general. normalize values to between 0 and 255 apply a gaussian blur normalize values to between 0 and 1 apply slight height restrictions based on distance to center of map ...

For 7DRL this year I developed a new map generation algorithm called Pigeon Hole Stepping. Fill the map with walls This map generation algorithm relies on negative building techniques. This concept will make more sense as the later steps are explained. Dig Hallways Essentially, we can assume that hallways that are parallel to each other are at least the length of a single room and up to two rooms apart in distance. Because of this assumption, we can make a random walker that will move around with this restriction in mind until map is filled completely. ...

Randomly generated dungeons are a lot of fun to program, but sometimes they can be pretty difficult to originally wrap your head around. For this years 7DRL I added Automata Generated Caverns as one of the algorithms I wanted to figure out completely. You can read more specifically on the theory here.