My Cellular Automata
These videos show 2 different rule-sets running on my 3-D cellular automaton based on cubic closed-packed array geometry. The vertical dimension (perpendicular to your display screen) is represented by color-coding (like looking down into a square aquarium from above).
Specifications
The vertical dimension is represented by color-coding (bright green signifies a vertical height closer to the viewer, while violet is furthest away). A colored cell blocks the view of other cells directly behind them along the line-of-sight. The resulting image is analogous to a color-coded contour map. The geometry of cell array is 3-dimensional Cubic Close-Packed, with 12 identical adjacent (touching) neighbors. All cells are equivalent with regards to geometry and rule application. The cell array is the above example is 70 x 70 x 70 cells. The size of the array is limited only by memory and execution speed. Larger arrays naturally update more slowly. Edges are wrapped in all 3 dimensions, corresponding to an overall 4-dimensional hyper-torus universe (I think). The initial "seed" population was a random distributions of several thousand "1" values (alive) sprinkled throughout a universe originally filled with "0" values (dead/empty), with an adjustable overall occurrence of "1" (typically about 5%). There are 35 distinct arrangements of cell-states around a central cell (counting reflections and/or rotations as equivalent). Each rule-set also takes into account the state of central cell, thus have 70 distinct possible cell patterns based on a central cell and its 12 neighbors. This is a non-totalistic cellular automaton (unlike Conway's totalistic Game of Life). Every rule-set is a 70-digit binary number (which can be also be displayed as a decimal number for convenience). There are 2^70 (= 1.18 x 10^21) different possible rule-sets, a very vast rule-space. Each digit in the binary rule-set is a "sub-rule" that specifies the state of the central cell in the next time-step (generation) when the corresponding arrangement of neighbors (assigned to the aforementioned digit) has been detected.
The rule-sets used above are given in the video descriptions. Most rule-sets (in the hundreds I have searched) are uninteresting, either dying or overpopulating quickly, or producing structureless chaos, or giving static ("still-life") or simple repeating ("blinker") patterns. A very small minority of rule-sets seem to provide examples of Wolfram's "Class IV" behavior, producing complex localized structures, sometimes long-lived. Simple "gliders" have been spotted in many rule-sets. I have barely "scratched the surface" in my exploration of different rule-sets, so I expect to find more interesting examples. The Cubic Close-Packed geometry of the cell array seems to work well, and will be maintained.
Return to Main Menu
automata.htm
updated 2021.12.14