I am trying to apply a convolutional deep Learning neural network to traffic jam data. I heard from different Web pages that neural networks weights and hyperparameters can be initialized with genetic algorithms. Can someone provide me with the tutorial or example on this topic, or code?

I am not much familiar with deep learning. I know about neural works but did not study much the novelty in the deep neural networks. But for neural networks in general, yes, you can use genetic or evolutionary algorithms to evolved some neural networks, so think it is certainly doable for deep learning. Someone has actually certainly tried this already so perhaps that you should do some search to find related papers first.