Feature-based aggregation and deep reinforcement learning: a survey and some new implementations
Feature-based aggregation and deep reinforcement learning: a survey and some new implementations is a scholarly work, published in 2019 in ''IEEE/CAA Journal of Automatica Sinica''. The main subjects of the publication include Markov chain, artificial intelligence, reinforcement learning, deep reinforcement learning, focus, implementation, mathematical optimization, biological function, artificial neural network, feature, Markov decision process, aggregate, adversarial machine learning, machine learning, and computer science. The authors introduce features of the states of the original problem, and authors formulate a smaller "aggregate" Markov decision problem, whose states relate to the features.