Genetic Algorithms and Machine Learning for Programmers

Did It Work?

You can run the genetic algorithm like this:

Boom/ga.py

def fire():
  epochs = 10
  items = 12
  height = 5
  width = 10

  generation = random_tries(items)
  generation0 = list(generation) # save to contrast with last epoch

  for i in range(1, epochs):
    results = []
    generation = crossover(generation, width)
    mutate(generation)

  display_start_and_finish(generation0, generation, height, width)

What you do in the display function is up to you. ​Plotting​ plots the arcs of the cannonballs from the first and last generation, so shows any improvement on the initial random attempts. There are other ways to assess your solutions. ​Counting​ discusses a few things worth counting at each epoch to see if the GA is learning. ...