Posted on Categories:CS代写, Machine Learning, 机器学习, 计算机代写

# CS代写|机器学习代写Machine Learning代考|KIT315 Introduction of Trafﬁﬁc Engineering

avatest™

## avatest™帮您通过考试

avatest™的各个学科专家已帮了学生顺利通过达上千场考试。我们保证您快速准时完成各时长和类型的考试，包括in class、take home、online、proctor。写手整理各样的资源来或按照您学校的资料教您，创造模拟试题，提供所有的问题例子，以保证您在真实考试中取得的通过率是85%以上。如果您有即将到来的每周、季考、期中或期末考试，我们都能帮助您！

•最快12小时交付

•200+ 英语母语导师

•70分以下全额退款

## CS代写|机器学习代写Machine Learning代考|Introduction of Trafﬁﬁc Engineering

Traffic Engineering (TE) is used to improve the overall Quality of Service (QoS) of the network. For a set of network flows ${ }^1$ with source and destination nodes, TE selects one or multiple paths to forward each flow for a specific objective (e.g., minimizing the maximum link utilization in the network) (Trimponias et al. 2019).
Classic TE solutions include Shortest Path First (SPF) (Open shortest path first 2019), which routes all flows through their shortest paths, and Equal-Cost Multipath Routing (ECMP), which evenly splits the volume of flows among multiple paths. These solutions are static and do not consider traffic pattern because routing rules are fixedly deployed regardless of traffic distribution in the network, and thus their performance is not optimal.

Designing an optimal TE solution based on the traffic distribution information is non-trivial (Trimponias et al. 2019). Existing solutions usually formulate the network as a static optimization problem using complicated analysis of the network traffic and solve the optimization problem with a specific objective to obtain the flow routing. However, these solutions have two major disadvantages. First, thoroughly and accurately understanding network traffic is very hard since network traffic can change dynamically. A traffic pattern-centric formulation can hardly cover all kinds of scenarios since different scenarios usually exhibit different traffic patterns. Second, the formulated optimization problems are usually NP-hard and thus hard to solve for obtaining optimal result (Karakus and Durresi 2017; Low and Lapsley 1999; Trimponias et al. 2019).

## CS代写|机器学习代写Machine Learning代考|Machine Learning-Based Solutions

Emerging Machine Learning (ML) techniques provide new opportunities to design good TE solutions. ML-based TE solutions analyze the network traffic and generate network policies without human experience. The ML-based TE solutions can be generally classified into three categories: supervised learning-based solutions, unsupervised learning-based solutions, and RL-based solutions. Supervised learning-based solutions mainly use various Deep Neural Networks (DNNs) to analyze network traffic. In this category, a certain neural network is designed and trained with a large amount of labeled network data to accurately extract required features (Huang et al. 2014; Kato et al. 2017; Parsaei et al. 2017). Unsupervised learning-based solutions employ unsupervised learning algorithms to analyze network traffic and extract traffic feature (Liu et al. 2007; Parsaei et al. 2017; Yan and Liu 2014). These first two categories can solve the first problem of TE as mentioned above by fetching the characteristics from the traffic. However, for the second problem, they still need manually designed heuristic algorithms to efficiently solve the formulated NP-hard problem. Besides, these two categories both have their own limitations. It is hard for supervised learning-based solutions to acquit a large amount of labeled data, and it is widely acknowledged that the accuracy of the feature extraction is limited for unsupervised learning-based solutions.

RL-based solutions can overcome the aforementioned disadvantages of supervised learning and unsupervised learning-based solutions (Mnih et al. 2015; Silver et al. 2016). In RL, an agent interacts with the network environment and takes the traffic distribution as the input to directly generate routing policies as the output (Boyan and Littman 1994; Lin et al. 2016). In this way, RL-based solutions can integrate traffic feature analyzing and routing policy generation into one procedure. However, RL usually uses a value table to establish the mapping between inputs and outputs and cannot handle large input and output space. To address this problem, DRL is proposed. DRL employs Deep Neural Networks (DNNs) in the agent to analyze the input data and trains the neural networks through the interaction between the agent and the network (Penghao et al. 2019; Xu et al. 2018; Yao et al. 2018). By taking advantage of the self-evolution ability of RL and high-dimensional data processing ability of the DNN, DRL shows great potential in generating control actions in networks.

Several recent works propose to use DRL for TE. For example, Yao et al. (2018) propose an SDN-based architecture named NetAI and use a simple example to illustrate the effectiveness of DRL for TE. TIDE (Penghao et al. 2019) uses DRL to automatically adjust the routing paths for networks, and DRL-TE (Xu et al. 2018) uses DRL to adaptively control the splitting ratio of each flow demand on candidate paths. All these works show advantages of DRL for TE in certain network scenarios.

## CS代写|机器学习代写Machine Learning代考|Machine Learning-Based Solutions

CS代写|机器学习代写Machine Learning代考 请认准UprivateTA™. UprivateTA™为您的留学生涯保驾护航。

## MATLAB代写

MATLAB 是一种用于技术计算的高性能语言。它将计算、可视化和编程集成在一个易于使用的环境中，其中问题和解决方案以熟悉的数学符号表示。典型用途包括：数学和计算算法开发建模、仿真和原型制作数据分析、探索和可视化科学和工程图形应用程序开发，包括图形用户界面构建MATLAB 是一个交互式系统，其基本数据元素是一个不需要维度的数组。这使您可以解决许多技术计算问题，尤其是那些具有矩阵和向量公式的问题，而只需用 C 或 Fortran 等标量非交互式语言编写程序所需的时间的一小部分。MATLAB 名称代表矩阵实验室。MATLAB 最初的编写目的是提供对由 LINPACK 和 EISPACK 项目开发的矩阵软件的轻松访问，这两个项目共同代表了矩阵计算软件的最新技术。MATLAB 经过多年的发展，得到了许多用户的投入。在大学环境中，它是数学、工程和科学入门和高级课程的标准教学工具。在工业领域，MATLAB 是高效研究、开发和分析的首选工具。MATLAB 具有一系列称为工具箱的特定于应用程序的解决方案。对于大多数 MATLAB 用户来说非常重要，工具箱允许您学习应用专业技术。工具箱是 MATLAB 函数（M 文件）的综合集合，可扩展 MATLAB 环境以解决特定类别的问题。可用工具箱的领域包括信号处理、控制系统、神经网络、模糊逻辑、小波、仿真等。