A reinforcement learning approach to improve the performance of the Avellaneda-Stoikov market-making algorithm | Publicación