Parallel Reinforcement Evolutionary Artificial Neural Networks (PREANN) is a framework of flexible multi-layer ANN’s with reinforcement learning based on genetic algorithms and a parallel implementation (using XMM registers and NVIDIA’s CUDA).