Loading...
Search for:
yousefizadeh--emadaddin
0.039 seconds
Some Model-free Discrete Reinforcement Learning Algorithms
, M.Sc. Thesis Sharif University of Technology ; Daneshgar, Amir (Supervisor)
Abstract
In this thesis, we review some methods related to model-free discrete reinforcement learning and their corresponding algorithms. Our main goal is to present existing methods in an integrated and formal setup, without compromising their mathematical accuracy or comprehensibility. We have done our best to fix the inconsistencies existing in notations and definitions appearing in different areas of the vast literature. We discuss dynamic programming methods, including policy iteration and value iteration and temporal difference methods as well as policy-based methods such as policy gradient, advantage actor-critic, TRPO, and PPO. Among value-based methods, we discuss Q-learning and C51 where we...
Evaluating the Energy Consumption of Fault-Tolerance Mechanisms In Processors Implemented on Sram-Based Fpgas
, M.Sc. Thesis Sharif University of Technology ; Ejlali, Alireza (Supervisor)
Abstract
With growing interest in the use of SRAM-based FPGAs in space and other radiation environments, there is a greater need for efficient and effective fault-tolerant design techniques specific to FPGAs. The soft errors vulnerability of SRAM-based FPGAs limits their usage in safety-critical applications. Moreover, the rate of multiple soft errors increases due to the feature size reduction. Hence, this issue becomes a challenge against reliability of the implemented circuit on SRAM-based FPGAs. Appealing to specifics such as low cost and re-configurability in SRAM based FPGAs provide this ability to change implemented design remotely. This advantage is not negligible in safety critical...
Comparison of Vibration and Ultrasonic Signal Performance in Predicting the Remaining Useful Life of Rolling Element Bearings
, M.Sc. Thesis Sharif University of Technology ; Behzad, Mehdi (Supervisor) ; Mohammadi, Somayeh (Supervisor)
Abstract
The prediction of Remaining Useful Life (RUL) has become a critical tool in enhancing the availability of rotating machinery. By knowing the remaining time before a bearing failure occurs during operation, significant reductions in unplanned production downtimes and associated costs can be achieved. In this study, by analyzing the time signals of vibration and ultrasonic data related to the bearing of a 4.1 MW electromotor from installation to replacement, and by investigating the behavior trends and correlation between extracted features, RMS has been ultimately identified as the health indicator in both vibration and ultrasonic techniques for monitoring the bearing’s condition. First,...