klena (also known as project-e)

Find the most efficient keyboard layout using the genetic algorithm.

How It Works?

Klena uses genetic algorithms' main principles.

• Create a population consisting of 10 randomly generated layouts from the gene pool (abcdefghijklmnopqrstuvwxyz[];',./<)
• Calculate a fitness score for each layout.
• Create 10 offspring from the two fittest layouts (the lower the fitness score, the better).
• Calculate the fitness score for the new generation and continue to the last step.

Calculating The Fitness Score

Fitness score calculation simulates key presses while looping through the dataset.

• First, it calculates the finger travel distance by using a hard-coded weighted graph (assuming 8 fingers at the home row).
• Then it checks for the biagrams and reduces the score accordingly.
• After that the key bias (according to the physical locations of the keys) and the same finger bias (repetitive usage of the same finger) is added.

Status

• Most features are implemented.
• GUI and the main logic is working.

TODOs

• Normalize and balance the fitness score.
• Reduce the dataset size by using proper dimension reduction techniques.
• Multi-core implementation of the fitness calculation function.
• Fix indentation and naming inconsistencies.
• Handle command line arguments.
• Write unit tests.
• Use proper logging methods.