January 2020
Intermediate to advanced
346 pages
9h 8m
English
Various dedicated algorithms have been developed for carrying out reinforcement learning tasks—Q-Learning, SARSA, and DQN, to name a few. However, since reinforcement learning tasks involve maximizing a long-term reward, we can think of them as optimization problems. As we have seen throughout this book, genetic algorithms can be used for solving optimization problems of various types. Therefore, genetic algorithms can be utilized for reinforcement learning as well, and in several different ways—two of them will be demonstrated in this chapter. In the first case, our genetic algorithm-based solution will directly provide the agent's optimal series of actions. In the second case, it will supply ...
Read now
Unlock full access