Development is in progress.
Currently implemented features:
- Flexible definition of genotype models with ability to use custom data and set up parameters for each gene individually.
- Customizable mutation and crossover operators which behave accordingly to the defined genotype model. Library includes one-point crossover operator and random value mutation and shift operators.
- All standard genetic algorithm parameters, like the breaking conditions, controlled population cap, fitness function definition.
- Selection is based on ranking groups to reduce the chance of getting into local extremum and keep diversity. The cutoff curve can be manually defined.
- Customizable logging and statistics facilities.
This is headers only library, so just copy into your project folder to get it working. Also CMake script is provided for classic installation as well as for development.
See examples in source directory (the API layer is in development, so things will change to conceal many unnecessary details).