Reinforcement Learning

In ancient games like chess or Go, the brightest players can improve by studying strategies produced by a machine. Robotic systems practice their movements autonomously. In arcade games, learning agents can reach superhuman levels within a few hours. How do these spectacular reinforcement learning algorithms work? With clear explanations and straightforward examples in Java and Greenfoot, you can grasp the principles of reinforcement learning and apply them in your own intelligent agents. Greenfoot and the hamster model are simple yet powerful educational tools designed to convey fundamental programming concepts. We will turn figures like the Java hamster into learning agents that explore their environment independently. The second edition includes new topics such as genetic algorithms and artificial curiosity, as well as explanations of current algorithms like A3C and PPO, along with corrections and revisions.