Reinforcement Learning Based on Active Learning Method - Computer Science > Artificial IntelligenceReportar como inadecuado




Reinforcement Learning Based on Active Learning Method - Computer Science > Artificial Intelligence - Descarga este documento en PDF. Documentación en PDF para descargar gratis. Disponible también para leer online.

Abstract: In this paper, a new reinforcement learning approach is proposed which isbased on a powerful concept named Active Learning Method ALM in modeling. ALMexpresses any multi-input-single-output system as a fuzzy combination of somesingle-input-singleoutput systems. The proposed method is an actor-criticsystem similar to Generalized Approximate Reasoning based Intelligent ControlGARIC structure to adapt the ALM by delayed reinforcement signals. Our systemuses Temporal Difference TD learning to model the behavior of useful actionsof a control system. The goodness of an action is modeled on Reward-Penalty-Plane. IDS planes will be updated according to this plane. It is shownthat the system can learn with a predefined fuzzy system or without it throughrandom actions.



Autor: Hesam Sagha, Saeed Bagheri Shouraki, Hosein Khasteh, Ali Akbar Kiaei

Fuente: https://arxiv.org/







Documentos relacionados