Design a Fault Tolerant Controller with SARSA method based on Reinforcement Learning

Design a Fault Tolerant Controller with SARSA method based on Reinforcement Learning

Mohammad Hasan Olyaei Torqabeh1 Ali Olyaei Torqabeh2 Amin Noori3

1) Faculty of Electrical Engineering, Sadjad University of Technology, Mashhad, Iran -
2) Department of Computer Engineering, Faculty of Engineering, Ferdowsi University of Mashhad, Mashhad, Iran -
3) Faculty of Electrical Engineering, Sadjad University of Technology, Mashhad, Iran -

IOI Code : XAHK-CFHDA
Publication : 2nd International Conference on New Horizons in the Engineering Science 9th August 2018 / Istanbul Turkey / Yildiz Technical University(eng2018.ir)
Abstract :
    This paper discusses the design of a fault-tolerant controller based on Reinforcement Learning. In each system, the probability of a fault is high. Usually, in sensitive systems, faults can have bad consequences, so today, scientists have designed different controllers to control the fault. In this paper, first we talk about reinforcement learning and then refer to it in the design of a fault-tolerant controller. After choosing the state, the action and the reward function, the controller is designed and the results indicate that the designed controller has caused the system to achieve the desired value and control the system fault.

Keywords : fault tolerant control reinforcement learning state action reward