Network Topology Information Research Articles

A Path-Server Traffic Scheduling Algorithm for Wireless Network Load based on SDN

At present, the research on flow scheduling optimization around SDN has become the focus of international attention. Based on the current traffic control algorithm, especially the shortcomings of traffic scheduling optimization, Path-Server Traffic Scheduling (PSTS) algorithm is proposed in this paper to solve problems. Based on the real-time monitoring of network on Ryu controller, network topology, network load and server load information were obtained to choose the optimal routing scheme, thus achieving the goal of improving the overall network performance. Through the interaction between the components, the work flow of the Ryu controller is designed, and the channel quality information among users is maintained by the link state management module. When the small base station receives the data packet sent by the SDN exchange price, it will temporarily store it in the data cache module. Then, according to the amount of cached data and the channel quality of each user, the optimal time slot of the wireless network resource allocation scheme is comprehensively determined, and the data packet is sent to the corresponding client. In the view of proposed design, the Mininet simulation platform is used to simulate the SDN network in this paper. Based on the simulation platform, the performance of the algorithm is analyzed and compared. Besides, bandwidth utilization, average transmission delay, system utility and terminal usage change, interrupt rate and terminal usage change of different algorithms in SDN wireless network are analyzed and compared through data. All experiments have proved that the research content proposed in this paper has an obvious effect on network load control, which shows network performance.

MRWMDA: A novel framework to infer miRNA-disease associations

MicroRNAs (miRNAs) are widely involved in a series of significant biological processes, which have been revealed and verified by accumulating experimental studies. The computational inference of the correlation between miRNAs and diseases is essential to facilitate the detection of disease biomarkers for disease diagnosis, prevention, treatment and prognosis. In this paper, a model with Multiple use of Random Walk with restart algorithm was introduced for the prediction of the MiRNA-Disease Association (MRWMDA). Based on diverse similarity measures, the model first implemented the random walk with restart (RWR) algorithm on the integrated similarity network to construct the topological similarity of miRNAs and diseases, which took full advantage of the network topology information. Then, the RWR algorithm was applied in the miRNA topological similarity network, and a steady probability of each miRNA-disease pair was obtained to prioritize miRNA candidates. In particular, the initial probability of the RWR algorithm was determined by utilizing the combination of the recommendation algorithm and the maximum similarity method. The proposed model achieved significant improvement in prediction compared with previous models, with an AUC of 0.9353 and an AUPR of 0.4809. In addition, case studies of breast neoplasms and lung neoplasms representing different disease types further demonstrated the excellent ability of MRWMDA in detecting potential disease-associated miRNAs. These performance analyses indicated that MRWMDA could be an effective and powerful biological computational tool in relevant biomedical studies.

Hi-GCN: A hierarchical graph convolution network for graph embedding learning of brain network and brain disorders prediction

PurposeRecently, brain connectivity networks have been used for the classification of neurological disorder, such as Autism Spectrum Disorders (ASD) or Alzheimer's disease (AD). Network analysis provides a new way for exploring the association between brain functional deficits and the underlying structural disruption related to brain disorders. Network embedding learning that aims to automatically learn low-dimensional representations for brain networks has drawn increasing attention in recent years. MethodIn this work we build upon graph neural network in order to learn useful representations for graph classification in an end-to-end fashion. Specifically, we propose a hierarchical GCN framework (called hi-GCN) to learn the graph feature embedding while considering the network topology information and subject's association at the same time. ResultsTo demonstrate the effectiveness of our approach, we evaluate the performance of the proposed method on the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset and Autism Brain Imaging Data Exchange (ABIDE) dataset. Extensive experiments on ABIDE and ADNI datasets have demonstrated competitive performance of the hi-GCN model. Specifically, we obtain an average accuracy of 73.1%/78.5% as well as AUC of 82.3%/86.5% on ABIDE/ADNI. The comprehensive experiments demonstrate that our hi-GCN is effective for graph classification with brain disorders diagnosis. ConclusionThe proposed hi-GCN method performs the graph embedding learning from a hierarchical perspective while considering the structure in individual brain network and the subject's correlation in the global population network, which can capture the most essential embedding features to improve the classification performance of disease diagnosis. Moreover, the proposed jointly optimizing strategy also achieves faster training and easier convergence than both the hi-GCN with pre-training and two-step supervision.

AligNet: alignment of protein-protein interaction networks

BackgroundAll molecular functions and biological processes are carried out by groups of proteins that interact with each other. Metaproteomic data continuously generates new proteins whose molecular functions and relations must be discovered. A widely accepted structure to model functional relations between proteins are protein-protein interaction networks (PPIN), and their analysis and alignment has become a key ingredient in the study and prediction of protein-protein interactions, protein function, and evolutionary conserved assembly pathways of protein complexes. Several PPIN aligners have been proposed, but attaining the right balance between network topology and biological information is one of the most difficult and key points in the design of any PPIN alignment algorithm.ResultsMotivated by the challenge of well-balanced and efficient algorithms, we have designed and implemented AligNet, a parameter-free pairwise PPIN alignment algorithm aimed at bridging the gap between topologically efficient and biologically meaningful matchings. A comparison of the results obtained with AligNet and with the best aligners shows that AligNet achieves indeed a good balance between topological and biological matching.ConclusionIn this paper we present AligNet, a new pairwise global PPIN aligner that produces biologically meaningful alignments, by achieving a good balance between structural matching and protein function conservation, and more efficient computations than state-of-the-art tools.

ClusterM: a scalable algorithm for computational prediction of conserved protein complexes across multiple protein interaction networks

BackgroundThe current computational methods on identifying conserved protein complexes across multiple Protein-Protein Interaction (PPI) networks suffer from the lack of explicit modeling of the desired topological properties within conserved protein complexes as well as their scalability.ResultsTo overcome those issues, we propose a scalable algorithm—ClusterM—for identifying conserved protein complexes across multiple PPI networks through the integration of network topology and protein sequence similarity information. ClusterM overcomes the computational barrier that existed in previous methods, where the complexity escalates exponentially when handling an increasing number of PPI networks; and it is able to detect conserved protein complexes with both topological separability and cohesive protein sequence conservation. On two independent compendiums of PPI networks from Saccharomyces cerevisiae (Sce, yeast), Drosophila melanogaster (Dme, fruit fly), Caenorhabditis elegans (Cel, worm), and Homo sapiens (Hsa, human), we demonstrate that ClusterM outperforms other state-of-the-art algorithms by a significant margin and is able to identify de novo conserved protein complexes across four species that are missed by existing algorithms.ConclusionsClusterM can better capture the desired topological property of a typical conserved protein complex, which is densely connected within the complex while being well-separated from the rest of the networks. Furthermore, our experiments have shown that ClusterM is highly scalable and efficient when analyzing multiple PPI networks.

Random Network Coding Enabled Routing Protocol in Unmanned Aerial Vehicle Networks

Unmanned aerial vehicles (UAVs) are becoming important communication infrastructures. One major challenge of communications with UAV networks is the routing protocol design. Due to the inherent characteristics (e.g., dynamic network topology and limited UAV device capabilities), it is difficult to directly apply existing routing protocols that utilize network topology information and routing path explorations. In this article, two novel routing protocols are designed based on random network coding (RNC) for a swarm UAV network, where UAVs operate cooperatively as a swarm, enabling efficient routing process. The first routing protocol utilizes the unique feature of RNC: Original packets can be decoded as long as an UAV accumulates sufficient generations. This property can be used to effectively expedite the underlying routing process. The second routing protocol further improves the efficiency where each forwarding UAV only needs to create a new generation rather than decoding original packets. Accordingly, the duration of each hop can be significantly reduced. Extensive simulations have been conducted to evaluate the performance of the designed routing protocols. The simulation results demonstrate that our designed routing protocols can effectively enhance the performance on both average transmission delay and delay violation probabilities compared to benchmark methods.

Early identification of technological convergence in numerical control machine tool: a deep learning approach

The importance of technology convergence of multidisciplinary knowledge has increased recently, and it is a crucial way to spur emerging technologies. Therefore, to understand and identify the technology convergence, which refers to the combination of two or more technological elements for a new system with new functions, is an important issue for both the researchers and the company directors. To identify and investigate the patterns of technology convergence, this research examines the numerical control machine tool, which has typical characteristics of technology convergence in recent years. Based on the numerical control machine tool related publications published between 1997 and 2019, we perform a deep learning approach based on Graph Neural Network model using publication citation network topology and text information together, to identify the technology convergence trajectory and to examine the dynamic role of corresponding technology sub-fields in the technology convergence. The results show that there was an obvious increase for the interdisciplinary citations from information technology to NC machine tool in recent years, and the technology convergence on NC machine tool is signal processing in machining and application of intelligent algorithms in motion control and process planning. In addition, the revelation of the technology convergence early identification contributes to the formation theory of emerging technologies that are interdisciplinary, and is of great interest to researchers, policy makers, and industrialists.

Distributed optimization via primal and dual decompositions for delay-constrained FANETs

This paper aims to optimize the different network parameters in a distributed manner for delay-constrained flying ad hoc networks (FANETs) without the global network topology information. To this end, each Unmanned Aerial Vehicle (UAV) calculates the average interference level during a certain time period to indicate the channel states. Next, we formulate the distributed optimization problem as a utility maximization problem, which jointly optimizes power control, rate allocation and delay-constrained routing. To obtain a distributed solution, a dual method is proposed to eliminate the link capacity constraint, and a primal decomposition method is employed to decouple the end-to-end delay constraint. Built on these two methods above, a distributed optimization algorithm is proposed in this work by considering the estimated one-hop delay for each transmission, which only uses the local channel information to optimize the sub-problems and limit the end-to-end delay. Finally, we deduce the relationship between the primal and dual solutions to underpin the advantage of the proposed algorithm. Experiments on simulate (and real) data demonstrate that the proposed algorithm effectively can improve network performances in terms of energy efficiency, packet timeout ratio and network throughput.

Graph Neural Networks with Information Anchors for Node Representation Learning

In the era of big data, the large-scale information network applications need to process and analyze increasingly complex graph structure relationships. However, traditional methods of representing network structures are difficult to reflect potential relationships between massive nodes. Graph Neural Network (GNN) based node representation learning is an emerging learning paradigm that embeds network nodes into a low dimensional vector space by retaining as much the information of network topology and node content as possible. However, existing GNN approaches ignore the distinction among the positions of nodes with similar topologies, which is usually crucial for many network prediction and classification tasks. In this paper, we propose a novel Graph Neural Network model based on information anchors, called A-GNN, where these anchors are defined as the important nodes that have a lot of interactive information with other ordinary nodes. In our model, the vectors obtained by the node representation learning contain the location information of ordinary nodes related to anchors. In A-GNN, we first designed the selection strategy of the set of anchors. Then the distance computation process was defined for any given target node to each anchor. Finally, we proposed the learning schema of a non-linear distance-weighted aggregation over the anchors. Therefore A-GNN can obtain global position information of all ordinary nodes relative to the anchors. Our proposed A-GNN is suitable for various network prediction tasks such as link prediction and node classification. We have conducted comparative experiments on five datasets. Experimental results show that A-GNN outperforms current state-of-the-art models.

Advanced spatial network metrics for cognitive management of 5G networks

The emerging fifth-generation (5G) mobile networks are empowered by softwarization and programmability, leading to the huge potentials of unprecedented flexibility and capability in cognitive network management such as self-reconfiguration and self-optimization. To help unlock such potentials, this paper proposes a novel framework that is able to monitor and calculate 5G network topological information in terms of advanced spatial metrics. These metrics, together with enabling and optimization algorithms, are purposely designed to address the complexity of 5G network topologies introduced by network virtualization and infrastructure sharing among operators (multi-tenancy). Consequently, this new framework, centred on a topology monitoring agent (TMA), enables on-demand 5G networks’ spatial knowledge and topological awareness required by 5G cognitive network management in making smart decisions in various autonomous network management tasks including but not limited to virtual network function placement strategies. The paper describes several technical use cases enabled by the proposed framework, including proactive cache allocation, computation offloading, node overloading alerting, and load balancing. Finally, a realistic 5G testbed is deployed with the central component TMA, together with the new spatial metrics and associated algorithms, implemented. Experimental results empirically validate the proposed approach and demonstrate the scalability and performance of the TMA component.

Deep Learning Architecture for Short-Term Passenger Flow Forecasting in Urban Rail Transit

Short-term passenger flow forecasting is an essential component in urban rail transit operation. Emerging deep learning models provide good insight into improving prediction precision. Therefore, we propose a deep learning architecture combining the residual network (ResNet), graph convolutional network (GCN), and long short-term memory (LSTM) (called “ResLSTM”) to forecast short-term passenger flow in urban rail transit on a network scale. First, improved methodologies of the ResNet, GCN, and attention LSTM models are presented. Then, the model architecture is proposed, wherein ResNet is used to capture deep abstract spatial correlations between subway stations, GCN is applied to extract network topology information, and attention LSTM is used to extract temporal correlations. The model architecture includes four branches for inflow, outflow, graph-network topology, as well as weather conditions and air quality. To the best of our knowledge, this is the first time that air-quality indicators have been taken into account, and their influences on prediction precision quantified. Finally, ResLSTM is applied to the Beijing subway using three time granularities (10, 15, and 30 min) to conduct short-term passenger flow forecasting. A comparison of the prediction performance of ResLSTM with those of many state-of-the-art models illustrates the advantages and robustness of ResLSTM. Moreover, a comparison of the prediction precisions obtained for time granularities of 10, 15, and 30 min indicates that prediction precision increases with increasing time granularity. This study can provide subway operators with insight into short-term passenger flow forecasting by leveraging deep learning models.

Density Visualization Pipeline: A Tool for Cellular and Network Density Visualization and Analysis.

Neuron classification is an important component in analyzing network structure and quantifying the effect of neuron topology on signal processing. Current quantification and classification approaches rely on morphology projection onto lower-dimensional spaces. In this paper a 3D visualization and quantification tool is presented. The Density Visualization Pipeline (DVP) computes, visualizes and quantifies the density distribution, i.e., the “mass” of interneurons. We use the DVP to characterize and classify a set of GABAergic interneurons. Classification of GABAergic interneurons is of crucial importance to understand on the one hand their various functions and on the other hand their ubiquitous appearance in the neocortex. 3D density map visualization and projection to the one-dimensional x, y, z subspaces show a clear distinction between the studied cells, based on these metrics. The DVP can be coupled to computational studies of the behavior of neurons and networks, in which network topology information is derived from DVP information. The DVP reads common neuromorphological file formats, e.g., Neurolucida XML files, NeuroMorpho.org SWC files and plain ASCII files. Full 3D visualization and projections of the density to 1D and 2D manifolds are supported by the DVP. All routines are embedded within the visual programming IDE VRL-Studio for Java which allows the definition and rapid modification of analysis workflows.

Application-Aware SDN-Based Iterative Reconfigurable Routing Protocol for Internet of Things (IoT).

The central intelligence offered by Software Defined Networking (SDN) promise the smart and reliable reconfiguration which enables the scalability of dynamic enterprise networks. The decoupled forwarding plane and control plane of SDN infrastructure is a key feature that supports the SDN controller to extract the physical network topology information at runtime to formulate network reconfigurations. This SDN-based network reconfiguration enables application-aware routing capability for Internet of Thing (IoT). However, these IoT enabled SDN-based routing protocols face some performance limitations in iterative reconfiguration process due to complete centralized path selection mechanism To this end, in this paper, we propose SDN-Based Application-aware Distributed adaptive Flow Iterative Reconfiguring (SADFIR) routing protocol. The proposed routing protocol enables the distributed SDN iterative solver controller to maintain the load-balancing between flow reconfiguration and flow allocation cost. In particular, the proposed routing protocol of SADFIR implements multiple SDN controllers that collaborate with network devices at forwarding plane to develop appropriate clustering strategy for routing the sensed information. This distributed SDN controllers are assisted to clustering topology that successfully map the residual network resources and also enable unique multi-hop application-aware data transmission. In addition, the proposed SADFIR monitor the iterative reconfiguration settings according to the network traffic of heterogeneity-aware network devices. The simulation experiments are conducted in comparison with the state-of-the-art routing protocols which demonstrates that SADFIR is heterogeneity-aware which is able to adopt the different scales of network with maximum network lifetime.

Identifying Drug Targets in Pancreatic Ductal Adenocarcinoma Through Machine Learning, Analyzing Biomolecular Networks, and Structural Modeling.

Pancreatic ductal adenocarcinoma (PDAC) is one of the leading causes of cancer-related death and has an extremely poor prognosis. Thus, identifying new disease-associated genes and targets for PDAC diagnosis and therapy is urgently needed. This requires investigations into the underlying molecular mechanisms of PDAC at both the systems and molecular levels. Herein, we developed a computational method of predicting cancer genes and anticancer drug targets that combined three independent expression microarray datasets of PDAC patients and protein-protein interaction data. First, Support Vector Machine–Recursive Feature Elimination was applied to the gene expression data to rank the differentially expressed genes (DEGs) between PDAC patients and controls. Then, protein-protein interaction networks were constructed based on the DEGs, and a new score comprising gene expression and network topological information was proposed to identify cancer genes. Finally, these genes were validated by “druggability” prediction, survival and common network analysis, and functional enrichment analysis. Furthermore, two integrins were screened to investigate their structures and dynamics as potential drug targets for PDAC. Collectively, 17 disease genes and some stroma-related pathways including extracellular matrix-receptor interactions were predicted to be potential drug targets and important pathways for treating PDAC. The protein-drug interactions and hinge sites predication of ITGAV and ITGA2 suggest potential drug binding residues in the Thigh domain. These findings provide new possibilities for targeted therapeutic interventions in PDAC, which may have further applications in other cancer types.

Robust Global Coordination of Networked Systems With Input Saturation and External Disturbances

This article refocuses on the global coordination of networked systems with input saturation and external disturbances. With the help of a novel backstepping approach, a multihop relay control algorithm subject to a group of modified heterogenous saturation functions is constructed, in which the number of the hop is equal to the order of the networked system. It is proved that these modified heterogenous saturation functions can develop a global unsaturated control algorithm for each agent. Then, some sufficient and necessary conditions are provided for the global consensus of the connected networked systems with input saturation in the absence of external disturbances. In the presence of the input saturation and external disturbances, a sufficient condition irrelevant to the interactive network topology information is proposed for the global swarm of connected networked systems. Finally, numerical simulations are provided to verify the theoretical results.

Two-phase optimization model for ride-sharing with transfers in short-notice evacuations

Optimization of on-demand transportation provisions and ride-sharing services in evacuations may provide increased network capacity and enhanced evacuation performance to transportation systems and improve equity and disaster preparedness for community and society. This paper proposes a two-phase model for optimizing trip planning and operations by integrating a ride-sharing process in short-notice evacuations, to allow a joint optimization of driver-rider matching and necessary transfer connections among shared vehicle trips. In the first phase, following network topology information and personal requests, a vehicle-space-time hyper dimensional network is developed by constructing vehicle-space-time vertexes and arcs. In the second phase, based on the constructed vehicle-space-time network, a new time-discretized multi-rider multi-driver network flow model is built to formulate ride-sharing with connecting transfers. A Lagrangian relaxation solution approach is designed to solve the model in a real-world network scenario. Numerical analyses are conducted with considerations given to the three operating parameters (detour tolerance of driver, penalty factor for transfer time, and maximum allowable parking time) in the method, and the analysis results show that the proposed model can not only meet the evacuation trip needs of the participating parties but it also supports personalized requests and on-demand accesses. A small sample network is used to theoretically test the whole model and the underlying concepts and solution strategy to show each step implemented in details, and finally the applicability of the method is demonstrated using the Chicago City network.

Integrating random walk with restart and k-Nearest Neighbor to identify novel circRNA-disease association

CircRNA is a special type of non-coding RNA, which is closely related to the occurrence and development of many complex human diseases. However, it is time-consuming and expensive to determine the circRNA-disease associations through experimental methods. Therefore, based on the existing databases, we propose a method named RWRKNN, which integrates the random walk with restart (RWR) and k-nearest neighbors (KNN) to predict the associations between circRNAs and diseases. Specifically, we apply RWR algorithm on weighting features with global network topology information, and employ KNN to classify based on features. Finally, the prediction scores of each circRNA-disease pair are obtained. As demonstrated by leave-one-out, 5-fold cross-validation and 10-fold cross-validation, RWRKNN achieves AUC values of 0.9297, 0.9333 and 0.9261, respectively. And case studies show that the circRNA-disease associations predicted by RWRKNN can be successfully demonstrated. In conclusion, RWRKNN is a useful method for predicting circRNA-disease associations.

Nonnegative matrix factorization for link prediction in directed complex networks using PageRank and asymmetric link clustering information

The aim of link prediction is to predict missing links in current networks or new links in future networks. Almost all the existing directed link prediction algorithms only take into account the links direction formation but ignored the abundant network topological information such as local and global structures. Therefore, how to preserve both local and global structure information is an important issue for directed link prediction. To solve this problem, in this paper, we are motivated to propose a novel Nonnegative Matrix Factorization via Asymmetric link clustering and PageRank model, namely NMF-AP. Specifically, we utilize the PageRank algorithm to calculate the influence score of the node, which captures the global network structure information. While we employ the asymmetric link clustering method to calculate the link clustering coefficient score, which preserves the local network structure information. By jointly optimizing them in the nonnegative matrix factorization model, our model can preserve both the local and global information at the same time. Besides, we provide an effective the multiplicative updating rules to learn the parameter of NMF-AP. Extensive experiments are conducted on ten real-world directed networks, experiment results demonstrate that the method NMF-AP outperforms state-of-the-art link prediction methods.

A Novel Network-Based Computational Model for Prediction of Essential Proteins

Identification of essential proteins is important for understanding cell survival and development, because even if only one of these proteins is missing, organisms cannot survive or develop. Since traditional methods for identifying essential proteins based on biological experiments are costly and inefficient, more and more computational models are proposed for predicting essential proteins in recent years. In this paper, a novel computational model called BSPM is proposed, in which, an original PPI network will be built based on known protein-protein associations first, and then topology information of the original PPI network will be adopted to measure the similarities between proteins based on the SimRank algorithm. Thereafter, a weighted PPI network can be obtained based on the similarities between proteins and the original PPI network. Finally, based on the weighted PPI network, the PageRank algorithm will be used to infer potential essential proteins. Moreover, in order to evaluate the performance of BSPM, we have compared the performance of BSPM with 14 classical prediction models in the field based on two different databases, and experimental results show that BSPM can achieve prediction accuracies of 92%, 81% and 76% out of the top 100, 200 and 300 candidate proteins separately, which not only are significantly better than those 14 competitive classical prediction models, but also means that BSPM can be used as an effective model for identifying essential proteins in the future.

Cognitive and Scalable Technique for Securing IoT Networks Against Malware Epidemics

The sheer volume of IoT networks being deployed today presents a major “attack surface” and poses significant security risks at a scale never encountered before. In other words, a single IoT device/node that gets infected with malware has the potential to spread the malicious activities across the network, eventually ceasing the network functionality or compromising the network. Simply detecting and quarantining the malware in IoT networks does not guarantee preventing malware propagation. On the other hand, use of traditional control theory for malware confinement is not effective, as most of the existing works do not consider real-time malware control strategies that can be implemented using uncertain infection information from the nodes in the network or have the containment problem decoupled from network performance. In response, in this work, we propose a two-pronged approach with malware detection at node-level, and confinement of malware at network-level. We deploy a recently proposed lightweight runtime malware detector at the node-level that employs Hardware Performance Counter (HPC) values for malware detection. This node-level malware information is combined with the malware propagation information and then fed during runtime to a stochastic predictive controller to confine the malware propagation without hampering the network performance. Synthesizing the node-level malware information with the model predictive containment strategy leads to achieving an average network throughput of nearly 200% of that of IoT network without any defense, and up to 160% of that of network with commonly employed state-of-the-art heuristic approaches for malware confinement. Furthermore, to scale with ever-increasing network topology sizes, we introduce a novel multi-attribute graph translation that can predict the network topology and node state information when provided with a snapshot of topology and node-level malware infection. The proposed multi-attribute graph translation has <5.88 Root Mean Square Error (RMSE) compared to the model predictive containment strategy and has shown nearly constant graph translation time and limited resource utilization independent of the network size.