Genetic algorithm optimizer using K-Means clustering with one way ANOVA algorithms
Use the package manager pip to install genetic_optimizer.
pip3 install genetic-optimizer
Package require =>Python 3.6.5, using genetic_optimizer in Python 2.x.x projects may not work correctly. If you are not sure about your python version, try:
import genetic_optimizer some_object = genetic_optimizer.Optimizer() some_object.generate(population_size=46, chromosome_size=8, equal_chromosomes=True, initialization_method='Random', representation='Binary', saving_method='csv')
These are default options for generator, you can change them in DEFAULTS.ini file in package directory
some_object.optimize(data=None, iterations=12, shuffle_scale=0.6, variety=0.8, chromosome_weight=0.0000001)
These are default options for optimizer, you can change them in STANDARDS.conf file in package directory
If optimizer got stuck at the beginning, that usually means that your dataframe with population is to small to find appropiate parents in order to create next generations.
For more description, please check out project wiki.
There is no mutation and crossover probability. Program matches parents according to group differences and create new child, built from the most successful pairs of genes from them. That means new generations are closer to each one and every child is not worse that genetic worse parent.
You can play off with options to achive more precised results.
- Fitness and parent selection process takes place recursively. If you are not using some RDD computing, program execution may be longer.
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
- micro editor
- Unix console
Code is mostly written in accordance with Google PEP style guide