Network congestion control based on neural network prediction algorithm

Chen Wei

Abstract:Neural Network is applied in network congestion control in this paper,then network congestion is predicted by Neural Network for network congestion avoidance and stable performance of network.Simulation results show the effectiveness of the algorithm.

Computer Science,Engineering

What problem does this paper attempt to address?

An Active Queue Management Scheme Using Neural Network Based Predictive Control

Li Wang,Shuxin Du,Jinguo Lin

DOI: https://doi.org/10.1109/iecon.2004.1432205

2004-01-01

Abstract:Active queue management (AQM) is a crucial and attractive theme in congestion control for IP network. Though control theory like Proportional Integral (PI) controller has been applied in AQM, the quality-of-service (QoS) of IP networks cannot always be guaranteed due to their nonlinear, time-varying and uncertain characteristics. In order to provide better QoS, a novel AQM algorithm, namely NNPC-AQM, is proposed based on predictive control, which requires less model accuracy. A predictor is constructed using two-layer linear neural network (NN) to predict the future queue length and a controller is composed of two-layer nonlinear NN to optimize the next drop probability. The performance of the proposed algorithm is verified and compared with PI controller by simulations. The results show that NNPC-AQM is robust against network parameters like number of TCP sessions and round trip time, and its performance has advantages over PI controller significantly.
Policy of Fuzzy Neural Network Based Congestion Control in High- Speed Network

Xiao-yan HE,Jie-yi WU,Guan-qun GU

DOI: https://doi.org/10.13328/j.cnki.jos.2001.01.006

2001-01-01

Journal of Software

Abstract:In this paper, a kind of traffic prediction a nd congestion control policy based on FNN (fuzzy neural network) is proposed for ATM (asynchronous transfer mode). Congestion control is one of the key problems in high-speed networks, such as ATM. Conventional traffic prediction method fo r congestion control using BPN (back propagation neural network) has suffered fr om long convergence time and dissatisfying precision, and it is not effective. T he fuzzy neural network scheme presented in this paper can solve these limitatio ns satisfactorily for its good capability of processing inaccurate information a nd learning. Finally, the performance of the scheme based on BPN is compared wit h the scheme based on FNN using simulations. The results show that the FNN schem e is effective.
A Rate Feedback Predictive Control Scheme Based On Neural Network And Control Theory For Autonomic Communication

Naixue Xiong,Athanasios V. Vasilakos,Laurence T. Yang,Fei Long,Lei Shu,Yingshu Li

DOI: https://doi.org/10.1007/978-0-387-09753-4_4

2009-01-01

Abstract:The main difficulty arising in designing an efficient congestion control scheme lies in the large propagation delay in data transfer which Usually leads to a mismatch between the network resources and the amount of admitted traffic. To attack this problem, this chapter describes a novel congestion control scheme that is based on a Back Propagation (BP) neural network technique. We consider a general computer communication model with multiple sources and one destination node. The dynamic buffer occupancy of the bottleneck node is predicted and controlled by using a BP neural network. The controlled best-effort traffic of the sources uses the bandwidth, which is left over by the guaranteed traffic. This control mechanism is shown to be able to avoid network congestion efficiently and to optimize the transfer performance both by the theoretic analyzing procedures and by the simulation studies.
Dynamic Traffic Congestion Simulation and Dissipation Control Based on Traffic Flow Theory Model and Neural Network Data Calibration Algorithm

Li Wang,Shimin Lin,Jingfeng Yang,Nanfeng Zhang,Ji Yang,Yong Li,Handong Zhou,Feng Yang,Zhifu Li

DOI: https://doi.org/10.1155/2017/5067145

IF: 2.3

2017-01-01

Complexity

Abstract:Traffic congestion is a common problem in many countries, especially in big cities. At present, China’s urban road traffic accidents occur frequently, the occurrence frequency is high, the accident causes traffic congestion, and accidents cause traffic congestion and vice versa. The occurrence of traffic accidents usually leads to the reduction of road traffic capacity and the formation of traffic bottlenecks, causing the traffic congestion. In this paper, the formation and propagation of traffic congestion are simulated by using the improved medium traffic model, and the control strategy of congestion dissipation is studied. From the point of view of quantitative traffic congestion, the paper provides the fact that the simulation platform of urban traffic integration is constructed, and a feasible data analysis, learning, and parameter calibration method based on RBF neural network is proposed, which is used to determine the corresponding decision support system. The simulation results prove that the control strategy proposed in this paper is effective and feasible. According to the temporal and spatial evolution of the paper, we can see that the network has been improved on the whole.

mathematics, interdisciplinary applications,multidisciplinary sciences
Predictive Control of Network Control Systems with Unknown Transmission Delay Via LSTM Network

Tao Chen,Lei Wang,Zhitao Liu,Hongye Su

DOI: https://doi.org/10.23919/ascc56756.2022.9828289

2022-01-01

Abstract:In networked control systems (NCSs), network delay is inevitable. The delay may degrade the control quality and even cause the control system to unstable. Prediction control is widely used to compensate for the network delay. In the controller-actuator channel, since the network delay of the packet to be transmitted is unknown, the controller always generates a long control sequence for safety, which causes a large amount of transmitted data in the network and thus increases the network burden. In this paper, to reduce the amount of the transmitted data, a long short-term memory (LSTM) neural network is first adopted to predict the network delay of the packet to be sent. Then the controller uses the predicted network delay to determine the prediction horizon. In this way, when the network delay is small, the controller only generates and transmits a short control sequence, and thus the amount of data transmitted can be reduced. In addition, the fixed-horizon and a vary-horizon predictive control are also given for comparison. Finally, simulation results are presented to verify the effectiveness and show the superiority of the proposed method.
A Novel Ad-hoc Network Congestion Forecast Method Based on Fuzzy Neural Network

LI Qian-Mu,QI Yong,XU Man-Wu,ZHANG Hong,LIU Feng-Yu

DOI: https://doi.org/10.3969/j.issn.1002-137X.2006.10.009

2006-01-01

Computer Science

Abstract:A Ad-hoc network congestion and rate control scheme based on fuzzy neural network(ACRC)is proposed. The normal technology of load balancing, such as master-slave scheduler and threshold scheduler, always assigns the task based on the present load of workstation. So, the resource utility is low and control of load balancing is lagged. The method can solve these limitations satisfactorily for its good capability of processing inaccurate information and self-learning. It not only mitigates congestion rapidly but also adjusts each sensor’s source rate to further reduce energy consumption. The results of simulations show that the ACRC method is effective.
A Short-term Traffic Flow Forecasting Method Based on Combination of Fuzzy C-mean Clustering and Neural Network

杨世坚,贺国光

DOI: https://doi.org/10.3969/j.issn.1001-4098.2004.08.020

IF: 2.034

2004-01-01

Systems Engineering

Abstract:Short-term traffic flow prediction is the basis of dynamic traffic control and guidance. In this article, a new fore- casting method, which combines fuzzy C-mean clustering and neural network techniques, is used to forecast short-term (traffic) flow. Based on the same real-time data, the forecasting results of this method are computed, and compared to other methods, e.g., Bp neural network and fuzzy neural network. The compute outcomes show that this method can produce (more) exact prediction results than other two methods.
Bandwidth prediction and congestion control for ABR traffic based on neural networks

Zhixin Liu,Xinping Guan,Huihua Wu

DOI: https://doi.org/10.1007/11760191_30

2006-01-01

Abstract:In this paper, queue length in the buffer is considered as a scale to evaluate the quality of service of communication network. To realize that the system is stable and avoid the data congestion under complex environments, the neural networks predictor and controller are designed, which can predict the bursty available bandwidth for ABR traffic effectively and force the queue level in the buffer to the desired region, respectively. The fairness of different connections is achieved through fair algorithm. The simulations verify the effectiveness of the proposed scheme.
Neural Congestion Control Algorithm in ATM Networks with Multiple Node

Ruijun Zhu,Fuliang Yin,Tianshuang Qiu

DOI: https://doi.org/10.1007/978-3-540-28648-6_47

2004-01-01

Abstract:In ATM networks, congestion control is a distributed algorithm to share network resources among competing users. It is important in situation where the availability of resources and the set of competing users vary over time unpredictably, round trip delay is uncertain and constraints on queue, rate and bandwidth are saturated, which results in wasted bandwidth and performance degradation. A neural congestion control algorithm is proposed by real-time scheduling between the self-tuning neural controller and the modified EFCI algorithm, which makes the closed-loop systems more stable and robust with respect to uncertainties and more fairness in resources allocation. Simulation results demonstrated the effectiveness of the proposed controller.
ROBUST CONGESTION CONTROL IN HIGH SPEED COMMUNICATION NETWORKS: A MODEL PREDICTIVE CONTROL APPROACH

Hua O. Wang,Yongru Gu,Huajing Fang

DOI: https://doi.org/10.3182/20020721-6-es-1901.00098

2002-01-01

IFAC Proceedings Volumes

Abstract:In this paper, the design of explicit rate-based congestion control in high speed communication networks is considered. At a bottleneck node, there are multiple best-effort sources competing with other high priority cross traffic sources. The goal of congestion control is to achieve high link utilization, low packet loss, low delay, and fairness among the best-effort sources. In this paper, the high priority traffic is described by an autoregressive integrated moving average (ARIMA) process. To deal with the propagation delays associated with the best-effort sources, model predictive control, particularly, generalized predictive control, techniques are proposed to solve the congestion problem here. It is demonstrated that the proposed controller performs well and is robust to delay uncertainties. In addition, in a multiple-nodes configuration, the controller provides max-min fairness.
ACCP: adaptive congestion control protocol in named data networking based on deep learning

Tingting Liu,Mingchuan Zhang,Junlong Zhu,Ruijuan Zheng,Ruoshui Liu,Qingtao Wu

DOI: https://doi.org/10.1007/s00521-018-3408-2

2018-01-01

Neural Computing and Applications

Abstract:Named data networking (NDN) is a novel network architecture which adopts a receiver-driven transport approach. However, NDN is the name-based routing and source uncontrollability, and network congestion is inevitable. In this paper, we propose an adaptive congestion control protocol (ACCP) which is divided into two phase to control network congestion before affecting network performance. In the first phase, we employ the time series prediction model based on deep learning to predict the source of congestion for each node. In the second phase, we estimate the level of network congestion by the average queue length based on the outcomes of first phase in each router and explicitly return it back to receiver, and then the receiver adjusts sending rate of Interest packets to realize congestion control. Simulation experiment results show that our proposed ACCP scheme has better performance than ICP and CHoPCoP in terms of the high utilization and minimal packet drop in a multi-source/multi-path environment.
Traffic Control Based on Dahlin Algorithm and Neural Network Prediction in Atm Networks

W Shen,R Feng,HH Shao

DOI: https://doi.org/10.1109/acc.2003.1239714

IF: 1.363

2012-01-01

Journal of Systems Engineering and Electronics

Abstract:The propagation delay in networks has a great adverse effect on rate-based traffic control. This paper proposes the composite control based on Dahlin algorithm feedback control and neural network feedforward predictive compensation on line for ABR communication in ATM networks, which can overcome the adverse effect caused by the delay on the control rapidity and stability better. The theoretical analysis and simulation research show that the scheme can make sources respond to the changes of network status rapidly, avoid the network congestion and utilize the link bandwidth better. Compared with PID control, cell loss rate is lower, link utilization rate is higher when buffer capacity is the same.
A Novel Predictive Congestion Control Algorithm for Urban Traffic Network

Ran Tao,Yugeng Xi,Dewei Li,Yiyao Luo

DOI: https://doi.org/10.23919/chicc.2017.8028959

2017-01-01

Abstract:For some big cities where congestion occurs frequently, this paper proposes a novel predictive congestion control algorithm, focusing on imposing hard constraints on links for avoiding congestion. Traditionally the urban traffic network is often optimized with some micro-integrated global performance indexes, which equalize the effect of congestion and the emergency of congestion may be ignored. However, for cities with heavy congestion, more attentions should be directly paid to congestion avoidance which may be particularly sensitive for urban traffic managers and individual drivers. We use the universal standard of length of the queue to describe the congestion status of the link. Through simplifying the macroscopic urban traffic S model, we approximately get a linear relationship between signals and queue lengths such that we can predict the future congestion status of the links. Based on this, we transform the traditional optimization into formulating the urban traffic congestion control as a constraint satisfaction problem(CSP) where congestion avoidance conditions are imposed on the links and the problem can be characterized by finding the feasible solution for a group of linear inequalities. If the problem is non-feasible, we further develop a hierarchical strategy to guarantee the most important links without congestion. Simulation results show that the control approach significantly decreases the number of congested links compared to other controllers.
A Chaotic Dynamic Congestion Prediction Method under the Space-Terrestrial Integrated Network.

Hua Qu,Changyu Wei,Xiaodong Yuan,Jihong Zhao

DOI: https://doi.org/10.1109/ICCCS55155.2022.9846732

2022-01-01

Abstract:The space-terrestrial integrated network is an important research field of the future network. The dynamic and heterogeneous nature of the space-terrestrial integrated network brings challenges to the research of congestion prediction methods. In this paper, we propose a chaotic dynamic congestion prediction method under the space-terrestrial integrated network, which solves the problems of insufficient dynamics and low accuracy of the existing congestion prediction methods for the space-terrestrial integrated network. The chaotic dynamic congestion prediction method is to predict the time series by using the GRU neural network model of the improved particle swarm algorithm to optimize the parameters after wavelet analysis. The experimental results show that the prediction accuracy of the chaotic dynamic congestion prediction method is higher, and it is more suitable for the space-terrestrial integrated network.
Dynamic Buffer Allocation Based on Traffic Prediction with FIR Neural Network

Jiacheng Ou,Yuanming Wu,Jiuhe Zhong

DOI: https://doi.org/10.1109/icfcc.2010.5497733

2010-01-01

Abstract:Analysis of network traffic has shown that traffic data exhibits high degree of self-similarity at different time scale. The feature of self-similarity indicates that there exists a predictive structure of the network traffic. In this paper, an improved finite-impulse-response neural network which utilizes momentum-modified temporal back-propagation algorithm is proposed for traffic prediction. Then the prediction results are used for dynamic buffer allocation to reduce packet loss in queuing systems. Simulation results show that the FIR neural network has good ability for traffic prediction, and the dynamic buffer allocation scheme based on traffic prediction with FIR neural network can effectively reduce packet loss rate.
Contagion Process Guided Cross-scale Spatio-Temporal Graph Neural Network for Traffic Congestion Prediction

Mudan Wang,Huan Yan,Huandong Wang,Yong Li,Depeng Jin

DOI: https://doi.org/10.1145/3589132.3625639

2023-01-01

Abstract:Frequent traffic congestion has a detrimental effect on our travel experience and the overall quality of urban life. Accurate prediction of traffic congestion plays a pivotal role in alleviating the congestion problem. However, existing traffic prediction approaches primarily focus on extracting its local changing patterns, overlooking the importance of incorporating global dynamic patterns. This presents three challenges: 1) Complicated spatial and temporal information exists in local (microscopic) traffic patterns; 2) The propagation and dissipation patterns of global (macroscopic) traffic congestion exhibit complex dynamics across time and space; 3) Modeling the interactions between macro and micro changing patterns of congestion remains unknown. In this paper, we present a novel framework for traffic congestion prediction that integrates microscopic and macroscopic cross-scale spatiotemporal modeling. Our approach utilizes contagion dynamics to characterize congestion propagation and recovery at the network-wide scale. Additionally, we employ a spatio-temporal graph neural network to capture local traffic patterns. A key contribution is the introduction of a differentiable micro-macro transformation mechanism, enabling the aggregation of microscopic states into macroscopic ones in a differentiable manner during model training. Further, we utilize the knowledge derived from macro contagion dynamics to constrain the micro traffic patterns by employing the physics-informed neural network. Experiments on three real-world datasets of traffic congestion demonstrate that our prediction model consistently outperforms the state-of-the-art baselines.
Traffic Flow Predictive Control Method of Urban Correlation Intersection Based on Neural Network

Fu Hui,Hongrun Gang,Juan Xu,Lunhui Xu

2008-01-01

Abstract:In order to improve running efficiency of entire urban network, aimed at character of urban traffic flow which is high nonlinear and real time change, a traffic flow predictive control method of urban correlation intersection based on neural network was proposed. The principle of the method is to minimize global vehicle queue length. Neural networks were used to predict traffic volume of intersections and were trained by error between real traffic volume and predicted value. According to the given method, the computation method of cycle, split and phase offset was presented. Moreover, a simulation experiment shows that using global vehicle queue length as performance index is effective to improve running efficiency of urban network.
A Neural Network Adaptive Controller for Explicit Congestion Control with Time Delay

Bo Yang,Xinping Guan

DOI: https://doi.org/10.1007/978-3-540-28648-6_11

2004-01-01

Abstract:This paper examines explicit rate congestion control for data networks. A neural network (NN) adaptive controller is developed to control traffic where sources regulate their transmission rates in response to the feedback information from network switches. Particularly, the queue length dynamics at a given switch is modeled as an unknown nonlinear discrete time system with cell propagation delay and bounded disturbances. To overcome the effects of delay an iterative transformation is introduced for the future queue length prediction. Then based on the causal form of the dynamics in buffer an adaptive NN controller is designed to regulate the queue length to track a desired value. The convergence of our scheme is derived mathematically. Finally, the performance of the proposed congestion control scheme is also evaluated in the presence of propagation delays and time-vary available bandwidth for robustness considerations.
A hybrid machine learning approach for congestion prediction and warning

Dongxue Li Yao Hu Chuliang Wu Wangyong Chen Feiyun Wang School of Mathematics and Statistics,Guizhou University,Guiyang,People's Republic of China

DOI: https://doi.org/10.1080/03081060.2024.2367751

2024-06-27

Transportation Planning and Technology

Abstract:Global traffic management encounters a significant challenge in traffic congestion. This paper presents a hybrid machine learning method for predicting traffic congestion. It leverages Convolutional Neural Network-Gated Recurrent Unit (CNN-GRU) network for parameter prediction and combines clustering and classification models, using the K-means algorithm to categorize traffic states into four levels and constructing a KNN classification model based on this segmentation. This results in the K-means-KNN model. Predicted parameters are inputted into the K-means-KNN model for congestion level prediction. Validation with real traffic flow data shows that the CNN-GRU network can capture spatiotemporal features more effectively. For instance, in traffic flow prediction, it reduces the MAPE by 7.39% and 51.14% compared to CNN and GRU, respectively. K-means-KNN excels in traffic state discrimination, achieving a congestion prediction accuracy of 91.8%. These results underscore the efficacy of the hybrid machine learning method in assessing and predicting urban traffic congestion.

transportation science & technology
Machine Learning Based Traffic Prediction and Congestion Control Algorithms in Software Defined Networks

Yanying Xu

DOI: https://doi.org/10.1109/iist62526.2024.00035

2024-03-04

Abstract:With the rise of Software Defined Networking (SDN) and the tremendous success of Machine Learning (ML) methods in classification, prediction, and control tasks, finding new traffic engineering techniques to adaptively and dynamically manage or route traffic in networks, ensuring service quality, and improving user experience quality has become a hot research topic in network research. SDN is a new type of network architecture with enormous growth potential. Its core idea is to separate forwarding and control, and concentrate intelligence on the controller for centralized control. This paper proposes an ML based traffic prediction and congestion control algorithm to address the problem of uneven load on the control plane of SDN architecture due to the complexity and variability of network traffic. Historical traffic data is collected from the SDN controller and feature extraction is performed. The extracted features are used to train the ML model and predict future traffic. Based on predicted traffic information and perceived congestion, the controller can dynamically adjust network parameters to avoid or alleviate congestion. The experimental results show that the algorithm proposed in this paper can achieve more efficient network traffic management and congestion control in SDN, thereby improving network performance and user experience.

Computer Science,Engineering

Network congestion control based on neural network prediction algorithm

An Active Queue Management Scheme Using Neural Network Based Predictive Control

Policy of Fuzzy Neural Network Based Congestion Control in High- Speed Network

A Rate Feedback Predictive Control Scheme Based On Neural Network And Control Theory For Autonomic Communication

Dynamic Traffic Congestion Simulation and Dissipation Control Based on Traffic Flow Theory Model and Neural Network Data Calibration Algorithm

Predictive Control of Network Control Systems with Unknown Transmission Delay Via LSTM Network

A Novel Ad-hoc Network Congestion Forecast Method Based on Fuzzy Neural Network

A Short-term Traffic Flow Forecasting Method Based on Combination of Fuzzy C-mean Clustering and Neural Network

Bandwidth prediction and congestion control for ABR traffic based on neural networks

Neural Congestion Control Algorithm in ATM Networks with Multiple Node

ROBUST CONGESTION CONTROL IN HIGH SPEED COMMUNICATION NETWORKS: A MODEL PREDICTIVE CONTROL APPROACH

ACCP: adaptive congestion control protocol in named data networking based on deep learning

Traffic Control Based on Dahlin Algorithm and Neural Network Prediction in Atm Networks

A Novel Predictive Congestion Control Algorithm for Urban Traffic Network

A Chaotic Dynamic Congestion Prediction Method under the Space-Terrestrial Integrated Network.

Dynamic Buffer Allocation Based on Traffic Prediction with FIR Neural Network

Contagion Process Guided Cross-scale Spatio-Temporal Graph Neural Network for Traffic Congestion Prediction

Traffic Flow Predictive Control Method of Urban Correlation Intersection Based on Neural Network

A Neural Network Adaptive Controller for Explicit Congestion Control with Time Delay

A hybrid machine learning approach for congestion prediction and warning

Machine Learning Based Traffic Prediction and Congestion Control Algorithms in Software Defined Networks