Optimizing Dynamic Pricing with Reinforcement Learning