4

My question is about the possibility to implement a custom fitness function in DEAP/Python in my Genetic Programming implementation.

After search and reading DEAP official documentation, i don't find anything about it, so, if one of you could help me, i appreciate it.

Thanks.

Cristiano Siebert
  • 57
  • 7

1 Answers1

2

Are you sure you need a custom fitness function?

It's a bit confusing, but you might be referring to a custom evaluation function. This should return a number that then the fitness function tries to maximize or minimize.

A great example is https://deap.readthedocs.io/en/master/examples/ga_onemax.html

In this tutorial, the standard maximizing fitness function is set up:

creator.create("FitnessMax", base.Fitness, weights=(1.0,))
creator.create("Individual", list, fitness=creator.FitnessMax)

followed by the "custom" evaluation function:

def evalOneMax(individual):
    return sum(individual),

which is then registered to the toolbox:

toolbox.register("evaluate", evalOneMax)
Tom Malkin
  • 2,134
  • 2
  • 19
  • 35
  • Wow! I had not realized this, really that is the right way, thank you! – Cristiano Siebert Nov 06 '18 at 12:10
  • I had the same problem. In EC evaluation function is commonly called the fitness function. I think that DEAP uses the evaluation function to map genotype to quality of measure. Then fitness function for comparing this measure. – Talos Jan 18 '22 at 11:05