A simulation using a cellular automaton system designed to explore evolution caused by emergent behaviors.
-
Tile-based world with cells representing organisms
-
Each cell has age and energy
-
Simple rules for energy harvesting, evolution, reproduction and death
-
Visualized evolution over time with step-by-step updates
-
Nuklear for the GUI
-
Microsoft Visual C++ Runtime (only if you're on Windows)
Linux (GCC):
g++ -std=c++23 -Wall -O2 main.cpp -o evolution-sim -I<SDL2 include path> \
-I<Nuklear include path> -L<lib path> -lSDL2 -lSDL2_image -lm
Windows (MSVC):
cl /std:c++latest /W4 /O2 main.cpp /Fe:evolution-sim.exe /I"<SDL2 include path>" ^
/I"<SDL2_image include path>" /I"<Nuklear include path>" /link Shell32.lib ^
/LIBPATH:"<SDL2 lib path>" SDL2main.lib SDL2.lib /LIBPATH:"<SDL2_image lib path>" ^
SDL2_image.lib /SUBSYSTEM:WINDOWS
-
Left click - select tile
-
Right click drag - camera panning
-
Mouse wheel - zoom
The world is a tilemap, where each tile can house a cell. Each cell has two basic properties: age and energy. Energy is the foundation of all life. When a cell runs out of energy, it dies.
At world generation, each tile is assigned a random energy source value, ranging from 0 to 75. The possible energy source values' probabilities are weighted so that lower values are more likely. Each tile also has a 10% (1 in 10) chance to house a live cell. Each initial live cell starts with a random amount of energy, ranging from 5 to 10.
Every generation, a live cell's age increases by 1. Every 10 generations, the energy of every tile in the world increases by 1.
If a live cell is on a tile with available energy, it will harvest up to 5 energy, with the harvested amount of energy being based on the following rules:
-
if the available energy is 100 or greater, the cell harvests 3 energy;
-
otherwise, if the available energy is 50 or greater, the cell harvests 2 energy;
-
otherwise, the cell harvests 1 energy;
-
if the cell is at least 40 generations old, it harvests an additional 2 energy;
-
otherwise, if the cell is at least 20 generations old, it harvests an additional 1 energy;
-
otherwise, the cell doesn't harvest any additional energy.
Of course, the actual harvested amount is limited by how much energy is available, and a cell can't harvest more energy than the tile below it can provide.
Every live cell's energy decreases by
Every cell that has at least 10 energy, is at least 10 generations old and has at least 1 free tile out of its 4 adjacent tiles has a 12.5% (1 in 8) chance to divide itself into two cells, splitting the mother cell's energy equally, and losing a unit of energy if the mother cell's energy is odd. When a cell divides itself, one daughter cell stays on the same tile as the mother cell, while the other is born on the most energy-rich free tile out of the 4 adjacent tiles to the original mother cell. The daughter cell that stays on the same tile inherits the mother cell's age, while the other one starts at 0 age. Both cells have a 50% (1 in 2) chance each for each evolution from the mother cell to inherit it.
If certain conditions are met, a cell has a chance to evolve to unlock a new ability/trait. Every evolution has to be a certain minimum age or above, requires a certain amount of energy and also takes a certain number of generations to complete, during which the cell cannot act - it cannot harvest more energy and cannot divide. Once a cell enters evolution, every generation of the evolution, a percentage of the total energy required for the evolution is subtracted from the cell's available energy. If the living cost causes the cell to run out of energy, it dies as usual, so every evolution attempt carries a risk. Every evolution also has a chance to be lost if unused in a generation, triggering regressive evolution.
Motility
When a cell has this evolution, if it has at least 3 energy after the living cost is applied, and there's a tile out of its 4 adjacent tiles that has more energy than its current tile, it will move to that tile, spending 1 energy in the process. If there are multiple adjacent tiles with more energy, it moves to the most energy-rich one.
- eligibility: 20 age
- cost: 20 energy
- timescale: 5 generations
- acquistion probability: 50% (1 in 2)
- loss probability: 4% (1 in 25)
Polydivision
When a cell has this evolution, the minimum energy required for its reproduction is doubled to 20, and its reproduction chance is halved to 6.25% (1 in 16). However, instead of dividing itself into two cells, it divides itself into as many cells as it can with the current provided space, i.e. the number of free adjacent tiles, allowing for it to divide itself into up to 5 cells without any additional energy costs (aside from its already doubled requirement). The energy of the mother cell is split equally between all new cells, losing all energy that can't be equally split. Only the daughter cell on the tile of the original mother cell retains the mother cell's age.
- eligibility: 40 age
- cost: 10 energy
- timescale: 10 generations
- acquistion probability: 25% (1 in 4)
- loss probability: 2.5% (1 in 40)
Energosynthesis
When a cell has this evolution, it has a 50% (1 in 2) chance to generate 1 energy every generation. If it has at least 1 free neighboring cell, it will always generate 1 energy. Additionally, if it has at least 4 free neighboring tiles, it generates an extra 1 energy, and if all 8 neighboring tiles are free, it generates another extra 1 energy, summing up to a total maximum of 3 energy per generation, if all neighboring tiles are free. The energy is generated before the living cost is applied.
- eligibility: 100 age
- cost: 30 energy
- timescale: 15 generations
- acquistion probability: 10% (1 in 10)
- loss probability: 12.5% (1 in 8)
A cell can only do one of the following during a single generation, and the priority is as listed:
- act "instinctively" - i.e. use an ability acquired through evolution
- divide itself
- initiate an evolution process
A cell dies when its energy reaches 0 after the living cost is applied. When a cell dies, the energy of the tile below it increases by the age of the now dead cell.
I'm working on this project solo. I was inspired by Conway's game of life, and decided to make something more complex and "alive", less order-based and more chaos-based. The entire thing is written in C++ with SDL2 and Nuklear.
-
More evolutions, as well as mutations, cell clusters, and generally more complex cell behavior
-
More simulation parameters
-
User intervention and control
-
Option to load and save simulation state
-
Cleaner GUI and a custom GUI implementation
-
Better rendering optimization