Genetic Algorithm Project

University of Campinas

Introduction to Artificial Intelligence

Group

• Aissa Hadj - 265189
• Lucas Zanco Ladeira - 188951
• Matheus Ferraroni - 212142
• Maria Vitória Rodrigues Oliveira - 262884
• Oscar Ciceri - 164786

File Tree

• Report - Report.
• Main - Main classes.
• IA - Maps to be utilized as input.
• Results - Results generated by the strategies.
• Plots - Plots of the results generated.

How to execute the files

There are two ways to execute the experiments:

1. Execute the python code called "main.py" and pass the parameters

The code considers the parameters to execute each strategy and map. It is possible to change what is the stop criteria, crossover type and more.

Run "python3 main.py --(strategy flag)" to execute.

For instance, "python main.py --population=50 --dataset=path --iteration_limit=100 --stop_criteria=1 --probs_type=0 --crossover_type=3 --crossover_rate=0.8 --mutation_type=0 --mutation_rate=0.03 --use_threads=0 --cut_half_pop=0 --replicate_best=0.1" executes a single feature selection with the mentioned parameters.

It is worth to point out that you may run "python main.py --help" to visualize all the possible flags and parameters.

2. Execute the shellcode called "execute.sh"

The code executes every configuration for each dataset and saves the results.

run "./execute.sh" and wait until it is done.

All results are saved in the results folder.

Good idea to use virtual env. Tested on Python 3.8