SARSA: Algorithm, Markov Decision Process, Reinforcement Learning, Machine Learning - Tapa blanda

SARSA

Taschenbuch. Condición: Neu. This item is printed on demand - Print on Demand Titel. Neuware -High Quality Content by WIKIPEDIA articles! SARSA(State-Action-Reward-State-Action) is an algorithm for learning a Markovdecision process policy, used in the reinforcement learning area ofmachine learning. It was introduced in the technical note 'OnlineQ-Learning using Connectionist Systems' by Rummery & Niranjan (1994)where the alternative name SARSA was only mentioned as a footnote. Thisname simply reflects the fact that the main function for updating theQ-value depends on the current state of the agent 'S1', the action theagent chooses 'A1', the reward 'R' the agent gets for choosing thisaction, the state 'S2' that the agent will now be in after taking thataction, and finally the next action 'A2' the agent will choose in itsnew state. Taking every letter in the quintuple (st , at , rt+1 , st+1at+1) yields the word SARSA.VDM Verlag, Dudweiler Landstraße 99, 66123 Saarbrücken 116 pp. Englisch. Nº de ref. del artículo: 9786131382338

Contactar al vendedor