.png){kind=icon}
返回
Deep Reinforcement Learning Task Scheduling Method for Real-Time Performance Awareness
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2025.3534980.png)
.png){kind=icon}
原文求助.png){kind=icon}
分享.png){kind=icon}
收藏
摘要
En 中文
Load balancing is essential for the efficient delivery of cloud computing services, ensuring stable operation and robust performance under high load conditions. However, existing load-balancing task scheduling algorithms struggle to adapt to load performance fluctuations in real-time, leading to inaccuracies in evaluating task execution efficiency and consequently impacting the quality of service in actual cloud task scheduling. To address this issue, we propose a real-time performance-aware task scheduling method based on the Soft Actor-Critic (RTPA-SAC) algorithm. This method dynamically detects server load performance changes in real-time, enhancing environmental consistency and adaptability in stochastic, dynamic task scheduling, thereby improving load balancing. First, we construct a bounded load performance loss function to evaluate task execution efficiency, considering the impact of parallel task interference. Next, a reward mechanism is introduced, which takes into account both load fluctuations and response times, optimizing task load variance within quality of service constraints to minimize response time. Finally, By leveraging the Soft Actor-Critic algorithm, the proposed scheduling strategy enhances exploratory and stable decision-making in task scheduling. Experimental results show that RTPA-SAC outperforms baseline methods in load balancing, evidenced by improvements in task response time, average task load variance, and task success rate.
Keyword:
Dynamic scheduling
Cloud computing
Heuristic algorithms
Scheduling
Load management
Real-time systems
Time factors
Stochastic processes
Servers
Load modeling
Task scheduling
load performance fluctuation
deep reinforcement learning
load balancing
期刊
IF:
3.6
论文数:
9.8W
被引数:
29.4W
机构
引用论文
.png){kind=icon}
Hypercalcaemia and diffuse osteolytic lesions in the acute phase of myeloid splenomegaly高钙血症和多发性溶骨性病变在髓性脾肿大急性期OAAI
OAAI分享收藏Multi-Objective Cloud Task Scheduling Optimization Based on Evolutionary Multi-Factor Algorithm
PREAI
分享收藏Intrusion detection in cloud computing based on time series anomalies utilizing machine learning基于时间序列异常的机器学习的云计算入侵检测OAAI
OAAI分享收藏Behavior and biology of the tomato psyllid, Bactericerca cockerelli, in response to the Mi‐1.2 gene
PREAI
分享收藏Cost-aware job scheduling for cloud inutances using deep reinforcement learning使用深度强化学习的云inutances成本感知作业调度
PREAI
分享收藏Batch Jobs Load Balancing Scheduling in Cloud Computing Using Distributional Reinforcement Learning
PREAI
分享收藏Work Scheduling in Cloud Network Based on Deep Q-LSTM Models for Efficient Resource Utilization
PREAI
分享收藏
查看更多内容