Reinforcement learning algorithms are only now starting to be applied in business settings — but could help companies solve complicated problems. Here’s how.

Most machine learning systems leverage historical data to make predictions. But learning through trial and error lead to more creative solutions.