Training neural networks to supervise genetic algorithms
Trains neural network to supervise genetic algorithms, the neural network will get informations about how the learning is going and make a choice about how to modify parameters such as the mutation rate, elitism, childs per survivors etc...
Currently, the network is trained on random instances from 5 different problems and take 2 to 5 times less simulations to get to the same results.