摘 要：
随着物联网和智能设备的快速发展，边缘计算作为一种新兴的计算范式，能够有效缓解云计算中心化架构带来的网络延迟和带宽压力问题，成为当前研究的热点领域。然而，边缘节点资源受限、网络环境动态变化以及任务负载不均衡等问题，对边缘计算系统的性能优化提出了严峻挑战。为此，本文聚焦于面向边缘计算的网络性能优化策略，旨在通过设计高效的资源分配与任务调度机制，提升系统整体性能并降低能耗。研究中提出了一种基于深度强化学习的自适应任务卸载算法，该算法能够根据实时网络状态和边缘节点资源情况，动态调整任务分配策略，从而实现任务执行时延与能耗之间的最佳权衡。同时，引入了分布式协同优化框架，以应对大规模边缘网络中的异构性和复杂性问题。实验结果表明，所提出的算法在多种典型场景下均表现出显著优势，相较于传统方法可将任务平均响应时间减少约30%，并将系统总能耗降低约25%。此外，本文还探讨了边缘节点协作机制对网络性能的影响，并验证了其在提高资源利用率和增强系统鲁棒性方面的有效性。综上所述，本研究不仅为边缘计算环境下的网络性能优化提供了新思路，也为未来智能化边缘计算系统的构建奠定了理论基础。
关键词：边缘计算；网络性能优化；深度强化学习；任务卸载；分布式协同优化
Optimization Strategies for Network Performance in Edge Computing
Abstract: With the rapid development of the Internet of Things (IoT) and smart devices, edge computing has emerged as a novel computing paradigm that can effectively alleviate the issues of network latency and bandwidth pressure caused by the centralized architecture of cloud computing, thus becoming a focal area of current research. However, challenges such as resource-constrained edge nodes, dynamically changing network environments, and imbalanced task loads pose significant obstacles to performance optimization in edge computing systems. To address these challenges, this study focuses on network performance optimization strategies for edge computing, aiming to enhance overall system performance and reduce energy consumption through the design of efficient resource allocation and task scheduling mechanisms. A deep reinforcement learning-based adaptive task offloading algorithm is proposed, which dynamically adjusts task distribution strategies according to real-time network conditions and resource availability at edge nodes, thereby achieving an optimal trade-off between task execution delay and energy consumption. Additionally, a distributed collaborative optimization fr amework is introduced to tackle heterogeneity and complexity issues in large-scale edge networks. Experimental results demonstrate that the proposed algorithm exhibits substantial advantages across various typical scenarios, reducing the average task response time by approximately 30% and lowering the total system energy consumption by about 25% compared to traditional methods. Furthermore, this study investigates the impact of edge node collaboration mechanisms on network performance and verifies their effectiveness in improving resource utilization and enhancing system robustness. In summary, this research not only provides new insights into network performance optimization in edge computing environments but also lays a theoretical foundation for the construction of future intelligent edge computing systems.
Keywords: Edge Computing; Network Performance Optimization; Deep Reinforcement Learning; Task Offloading; Distributed Collaborative Optimization
目 录
1绪论 1
1.1边缘计算与网络性能优化背景 1
1.2研究意义与价值分析 1
1.3国内外研究现状综述 1
1.4本文研究方法与技术路线 2
2边缘计算中的延迟优化策略 2
2.1延迟优化的关键挑战 2
2.2数据传输路径优化方法 3
2.3边缘节点部署策略研究 3
2.4实时任务调度机制设计 4
2.5延迟优化案例分析 4
3资源分配与能耗优化策略 5
3.1边缘计算资源分配问题定义 5
3.2动态资源分配算法研究 5
3.3能耗优化模型构建 6
3.4分布式计算任务分配方案 6
3.5资源与能耗优化实验验证 7
4安全性与可靠性保障策略 7
4.1边缘计算中的安全威胁分析 7
4.2数据隐私保护机制设计 8
4.3可靠性评估与增强方法 8
4.4安全通信协议优化策略 9
4.5安全与可靠性综合案例研究 9
结论 10
参考文献 11
致 谢 12
