Abstract
Genetic algorithms are algorithms that are inspired by the biological process of natural selection and belong to a larger class of algorithms called evolutionary algorithms (Carr, 2014). The genetic algorithms used in this project rely on changing the crossover types, which means that the algorithm changes the way that the matrix randomly mutates during the process of finding the ideal solution (Umbarkar and Sheth, 2015). The purpose of this experiment was to determine which genetic algorithm would sort a dataset in numerical order the most quickly. The independent variable (IV) in this experiment was the different types of the genetic algorithm used to sort the data in numerical order. The dependent variable was the time spent on the algorithms to sort through the dataset. The four levels of the IV were single-point crossover, two-point crossover, uniform crossover, and flat crossover. It was hypothesized that the single point crossover algorithm would sort the dataset the most quickly. First, the materials (algorithms, dataset, and code to run the algorithms) were obtained from Github user “dawidkopczyk.” Then, the algorithms were run with the modified code, and the data was collected and analyzed. The experiment showed that single-point crossover was the fastest, then two-point, uniform, and finally flat crossover. The hypothesis was proven correct due to the fact that single-point crossover requires the least amount of changes in-between steps, which is less work for the computer. Many papers support the results found in this experiment such as one written by Alden H. Wright. For further research, the sizes of the dataset could be changed. An improvement that could be made would be to use a more diverse dataset. This could lead to a more balanced outcome.