I've been using back-propagation to optimize neural networks without problems. But I've read in some books that genetic algorithms can be used to optimize an ANN. I want to know if its possible to use them as a learning algorithm.


1 Answer 1


Yes! There are essentially 2 ways of doing that: with fixed topology or evolving topology.

Fixed topology is the easiest. Just encode all weights as values in your chromossomes and evolve. You can use the network error as your fitness (to minimize, in this case). The disadvantage of this method is that you need to chose the network architecture manually. You can see this approach in action here.

But there is also the possibility of evolving the structure together with the weights. The most famous algorithm for this is NEAT. You can see it in action here and here.

  • $\begingroup$ Thank you bro!! thats what i nedd. Do you speak spanish by the way?? $\endgroup$ Aug 3, 2015 at 6:31
  • $\begingroup$ Just portuguese! At most portunhol hehe $\endgroup$
    – rcpinto
    Aug 3, 2015 at 14:54
  • $\begingroup$ is it possible compare the performance of these two learning algorithms(Back-propagation and Genetic Algorithms)?What are the parameters that i can compare?? $\endgroup$ Aug 4, 2015 at 15:34
  • $\begingroup$ Yes, it is. It depends on your problem. For instance, if it is a classification task, you could compare the generalization accuracy of the best model obtained from each technique. You could also compare the training time or memory. $\endgroup$
    – rcpinto
    Aug 4, 2015 at 15:43
  • $\begingroup$ Thank you men, you are helping me a lot, And for prediction problem what parameter should i do?? $\endgroup$ Aug 4, 2015 at 16:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.