Loading...
Search
| Friend's email | |
| Your name | |
| Your email | |
| enter code | |
This page was sent successfuly
Digital Object List
Bookmark
Review List
187 viewed
مدیریت انرژی در سامانه های تحملپذیر اشکال بحرانی - مختلط با استفاده از یادگیری ماشین
تقوی نیا، محمد حسین Taghavi Nia, Mohammad Hossein
Energy Management in Fault-Tolerant Mixed-Criticality Systems Using Machine Learning
Taghavi Nia, Mohammad Hossein | 2024
187
Viewed
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 56954 (19)
- University: Sharif University of Technology
- Department: Computer Engineering
- Advisor(s): Hessabi, Shaahin
- Abstract:
- Mixed-critical systems (MCSs) use a common hardware platform to schedule and execute tasks with multiple critical levels. Multi-core processors are commonly used by these systems, and batteries are frequently employed to provide the energy they require due to their portable nature or lack of access to large energy sources. Therefore, it is essential to use energy-management methods, such as dynamic voltage and frequency scaling (DVFS) and dynamic power management (DPM), to reduce their energy consumption. Using power/energy management methods can lead to missed deadlines due to increased execution time of tasks. In two-level mixed-critical systems, the lack of correct and timely execution of high-critical tasks will result in severe damage; Therefore, guaranteeing the reliability of high-critical tasks is very important in these systems. Usually, the inherent redundancy in the system is used so that the system is tolerant of problems that occur in high-critical tasks and provides the level of reliability required by them. Low-critical tasks do not need to guarantee a high level of reliability; However, it is not acceptable in today's applications to decrease the quality of service (QoS) level too much. In this research, for the first time, we have used deep reinforcement learning (DRL) to manage the energy of tasks in mixed-criticality systems. Also, the system is designed in such a way that tasks are executed before their deadline; it can tolerate problems that occur during high-critical tasks in the offline phase, ensure that the minimum level of quality of service for low-critical tasks is met, and increase this amount in the online phase. We have used the ER-POED and EDF algorithms to schedule periodic tasks, along with the hot standby-sparing fault tolerance method. The simulation results indicate that the designed system decreases energy consumption by 56% and, on average, by 37% compared to the baseline
- Keywords:
- Mixed-Criticality Scheduling ; Energy Optimization ; Deep Reinforcement Learning ; Multi-Core Platforms ; Fault Tolerance ; Service Quality ; Energy Management ; Machine Learning
-
محتواي کتاب
- view