Network Operations Research Articles

Improved GAN with fact forcing for mobility prediction

Mobility prediction based on users movement history is a prerequisite for proactive mobility and several other management operations in wireless networks. Recent Deep Learning (DL) studies for mobility prediction have shown improved accuracies using variety of different models that have high complexity and massive training data requirements. These limitations restrict the viability of mobility prediction driven proactive management of wireless networks. This paper overcomes the mentioned limitations by proposing Improved Generative Adversarial Network with Fact Forcing (iGAN-FF) that enhances the existing improved GAN (iGAN). iGAN-FF architecture consists of generator and discriminator neural networks, where generator predicts mobility (next Point of Attachment (PoA)) and discriminator classifies between the predicted PoA and the ground truth in adversarial learning. The adversarial learning reduces the needed amount of training data and enables use of simplistic models, and novel compounded fact forcing and feature matching method achieves higher accuracy. Moreover, the effects of data representation on performance are evaluated by transforming the Campus Mobility Dataset (CMD) and Operator Mobility Dataset (OMD) to One-hot, Binary, and Embedded vectors representations. The results confirm the merits of iGAN-FF as it outperforms iGAN, GAN-FF, GAN, LSTM based Next PoA (LNP), and Transformer based Next PoA (TNP). In particular, the highest accuracies of 99.57% and 83.3% are achieved with CMD One-hot and OMD Embedded representations, respectively. Notably, iGAN-FF shows robustness under data scarcity and secures ∼98% accuracy with only 20% of CMD data.

Channel pruning based on convolutional neural network sensitivity

Pruning is a useful technique for decreasing the memory consumption and floating point operations (FLOPs) of deep convolutional neural network (CNN) models. Nevertheless, at modest pruning levels, current structured pruning approaches often lead to considerable declines in accuracy. Furthermore, existing approaches often treat pruning rates as super parameters, neglecting the sensitivity of different convolution layers. In this study, we propose a novel sensitivity-based method for channel pruning that utilizes second-order sensitivity as a criterion. The essential concept is to prune insensitive filters while retaining sensitive ones. We quantify the sensitivity of the filter using the sum of the sensitivities of all weights in the filter, rather than the magnitude-based metric frequently applied in the literature. Furthermore, a layer sensitivity approach based on the Hessian eigenvalues of each layer is introduced into the process of automatically choosing the most appropriate pruning rate for each layer. Experiments on a variety of modern CNN architectures demonstrate that we can considerably enhance the pruning rate while sacrificing a small amount of accuracy, resulting in a reduction of more than 60% in FLOPs on CIFAR-10. Notably, on ImageNet, pruning based on ResNet50 decreased the FLOPs by 56.3% while losing only 0.92% of accuracy.

PROFITABILITY ANALYSIS OF PUBLIC SECTOR BANKS IN INDIA

Economic development of any nation depends on its efficient financial operations of the banking networks. Indian banking sector facing severe NPA issues. In order to overcome this problem, Government of India started merging activities. Some major mergers took place in recent years which made the Indian Banking Sector effective, innovative and competitive in the global level. Public sector banks which left as an independent public sector bank after merging process conducted by government of India in the last few years are taken as a sample for the study. The purpose of this article is to determine the financial factors affecting the performance of public sector banks in India. This article analyzed the financial data of six independent public sector banks for the period of four years from 2017-18 to 2020-21. For analysis purpose, used secondary data from various financial websites and only quantitative data is considered and ratio analysis and descriptive statistics are used for analysis purpose. The present study has examined the financial performance in the public sector banks which has a significant result in terms of Net Profit Margin, Return on Net Worth, Operating Expense / Total Income and Capital Adequacy Ratios Bank of India performed better. Central Bank of India functioned well in getting significant long term returns and all selected sample banks shown negative outcome in the relation of Net Profit / Total Funds. Bank of Maharashtra revealed its inefficiency in maintaining negative loan turnover ratio, while Central Bank of India more efficient bank which has Positive Loan turnover ratio. Outperformed Punjab and Sind Bank in terms of Total Assets Turnover Ratio, Advances / Loans Funds and Earnings Per Share. All selected six sample Banks has less than one (<1) Current ratio in shown inefficient in making its dues in a time. Punjab and Sind bank and UCO Bank used its huge amount of interest earned income for making payment of Interest expenses. From the study it is clear that there is a need to improve the overall profitable performance of selected sample of public sector banks by the government which is in need for its survival. KEYWORDS: Profitable Performance, Public sector banks, ratio analysis.

Learning to run a power network with trust

Artificial agents are promising for real-time power network operations, particularly, to compute remedial actions for congestion management. However, due to high reliability requirements, purely autonomous agents will not be deployed any time soon and operators will be in charge of taking action for the foreseeable future. Aiming at designing assistant for operators, we instead consider humans in the loop and propose an original formulation. We first advance an agent with the ability to send to the operator alarms ahead of time when the proposed actions are of low confidence. We further model the operator’s available attention as a budget that decreases when alarms are sent. We present the design and results of our competition “Learning to run a power network with trust” in which we evaluate our formulation and benchmark the ability of submitted agents to send relevant alarms while operating the network to their best.

A population-based incremental learning algorithm to identify optimal location of left-turn restrictions in urban grid networks

The treatment of left turns at signalized intersections drives the development of signal phasing and timing plans and also plays an important role in overall traffic network operations. Accommodating left turns allows for the most direct routeing but reduces intersection capacity, whereas restricting left turns improves capacity but requires some vehicles to travel longer distances. This paper proposes a population-based incremental learning (PBIL) algorithm to determine at which intersections left-turn restrictions should be enacted to maximize a network's operational performance. The performance of each configuration is tested in a micro-simulation environment on both perfect and imperfect square grid networks. Comparison with a partial enumeration of feasible options reveals that the PBIL algorithm is effective at identifying a near-optimal configuration of left-turn restrictions. The resulting configurations suggest that left turns should be generally restricted at intersections that carry the most flow. These intersections typically occur in the central portion of the network when demands are relatively uniform. Doing so helps to provide additional intersection capacity at the locations where it is most necessary, while minimizing the additional travel distance that is incurred due to detours caused by the left-turn restrictions. These provide insight as to how urban traffic networks might be managed to improve network efficiency by only enacting left-turn restrictions at a subset of locations.

Protecting infrastructure performance from disinformation attacks

Disinformation campaigns are prevalent, affecting vaccination coverage, creating uncertainty in election results, and causing supply chain disruptions, among others. Unfortunately, the problems of misinformation and disinformation are exacerbated due to the wide availability of online platforms and social networks. Naturally, these emerging disinformation networks could lead users to engage with critical infrastructure systems in harmful ways, leading to broader adverse impacts. One such example involves the spread of false pricing information, which causes drastic and sudden changes in user commodity consumption behavior, leading to shortages. Given this, it is critical to address the following related questions: (i) How can we monitor the evolution of disinformation dissemination and its projected impacts on commodity consumption? (ii) What effects do the mitigation efforts of human intermediaries have on the performance of the infrastructure network subject to disinformation campaigns? (iii) How can we manage infrastructure network operations and counter disinformation in concert to avoid shortages and satisfy user demands? To answer these questions, we develop a hybrid approach that integrates an epidemiological model of disinformation spread (based on a susceptible-infectious-recovered model, or SIR) with an efficient mixed-integer programming optimization model for infrastructure network performance. The goal of the optimization model is to determine the best protection and response actions against disinformation to minimize the general shortage of commodities at different nodes over time. The proposed model is illustrated with a case study involving a subset of the western US interconnection grid located in Los Angeles County in California.

Multi-hop quantum operation teleportation via different high-level partially entangled channels

To implement remote control operations in quantum networks, two schemes for the multi-hop quantum operation teleportation are proposed by using different partially entangled states. One is to adopt high-level partially Bell states as quantum channels, the other is to use high-level partially GHZ-Bell states. By comparing the two schemes, it is found that although both schemes have the same communication delay, the latter has less classical information cost and higher success probability of the communication.

MGM-4-FL: Combining federated learning and model gossiping in WSNs

Application of Federated learning (FL) as it is, in Wireless Sensor Networks (WSNs) in impractical. In fact, it requires the transmission of a large number of model parameters. Also, it implies the consumption of a massive amount of energy and communication resources at nodes located closer to the network elements that aggregate the models calculated by the federated learners. This is a well known problem in WSN called funneling effect. Solutions have been recently proposed that decentralize FL operations further by exploiting multihop communications. Nevertheless, the proposed approaches do not focus on the networking operations needed to support such decentralization in an efficient manner. Objective of this paper is to fill this gap by proposing an integrated networking/learning scheme, named Model Gossiping Method for FL, that supports distributed FL efficiently in wireless sensor networks.

General polynomial chaos in the current–voltage formulation of the optimal power flow problem

Mathematical optimization techniques play a key role in enabling the power system transition to sustainable energy and are used for a variety of applications such as scenario analysis, optimal planning and operational decision making. Power flow optimization, a.k.a., optimal power flow, is a building block for many applications in network operations and planning. This paper discusses the treatment of general polynomial chaos expansion for the current–voltage formulation of the optimal power flow problem. The power flow equations of the current–voltage formulation are linear, making their Galerkin projection significantly more tractable compared to formulations in the power–voltage space, while still being exact. Furthermore, auxiliary variables and quadratic constraints enable chance constraints as second-order-cone constraints. An additional advantage of this approach is that the Galerkin projection of the quadratic constraints is significantly less complex compared to those of non-linear constraints with a polynomial degree higher than two, as would be needed for expressing the original variables’ variance without the auxiliary variables. On average, the current–voltage formulation using auxiliary variables shows more than an order of magnitude speed-up with respect to the power–voltage formulation without auxiliary variables.

Batch reinforcement learning for network-safe demand response in unknown electric grids

We formulate a batch reinforcement learning-based demand response approach to prevent distribution network constraint violations in unknown grids. We use the fitted Q-iteration to compute a network-safe policy from historical measurements for thermostatically controlled load aggregations providing frequency regulation. We test our approach in a numerical case study based on real load profiles from Austin, TX. We compare our approach’s performance to a greedy, grid-aware approach and a standard, grid-agnostic approach. The average tracking root mean square error is 0.0932 for our approach, and 0.0600 and 0.0614 for, respectively, the grid-aware and grid-agnostic implementations. Our numerical case study shows that our approach leads to a 95% reduction, on average, in the total number of rounds with at least a constraint violation when compared to the grid-agnostic approach. Working under limited information, our approach thus offers lower but acceptable setpoint tracking performance while ensuring safer distribution network operations.

Complexity optimal air traffic assignment in multi-layer transport network for Urban Air Mobility operations

Large numbers of Urban Air Mobility (UAM) vehicles are expected to operate in urban airspace in the near future, exceeding the capacities of current airspace and Air Traffic Management (ATM) systems. This paper presents an air traffic assignment framework for 3D air transport networks in urban airspace to enable UAM operations at future demand levels. The individual vehicle dynamics are aggregated to describe the overall flow feature in this macroscopic model. Firstly, UAM operations are modeled as flows and structured in a three-dimensional two-way air transport network. Then, a complexity optimal air traffic assignment in urban airspace is formulated as an optimization problem. Based upon the Linear Dynamical System (LDS), a novel complexity metric is defined as objective function, which takes into account dynamic flow structure, congestion, and operational efficiency. A two-phase approach combining Simulated Annealing (SA) and Dafermos’ Algorithm (DA) is introduced to efficiently solve this problem. To validate the proposed model, a case study of an air transport network in Singapore’s urban airspace with two different demands is conducted. Comparative studies are carried out between the proposed algorithm and other widely used traffic assignment algorithms. The results show that the proposed approach is capable of assigning flows in an efficient and effective manner, reducing the complexity of the air transport network significantly. The results also show that optimizing the flow pattern reduces total complexity by 90.44%±0.53% and 92.12%±0.35% with 95% confidence interval, respectively in two scenarios. The framework may be useful for Air Navigation Service Providers (ANSP) in strategic planning for UAM operations and urban airspace design.

Research on Network Attack Traffic Detection HybridAlgorithm Based on UMAP-RF

Network attack traffic detection plays a crucial role in protecting network operations and services. To accurately detect malicious traffic on the internet, this paper designs a hybrid algorithm UMAP-RF for both binary and multiclassification network attack detection tasks. First, the network traffic data are dimensioned down with UMAP algorithm. The random forest algorithm is improved based on parameter optimization, and the improved random forest algorithm is used to classify the network traffic data, distinguishing normal data from abnormal data and classifying nine different types of network attacks from the abnormal data. Experimental results on the UNSW-NB15 dataset, which are significant improvements compared to traditional machine-learning methods, show that the UMAP-RF hybrid model can perform network attack traffic detection effectively, with accuracy and recall rates of 92.6% and 91%, respectively.

Analysis and Comparison of Different Approaches to Implementing a Network-Based Parallel Data Processing Algorithm

It is well known that network-based parallel data processing algorithms are well suited to implementation in reconfigurable hardware recurring to either Field-Programmable Gate Arrays (FPGA) or Programmable Systems-on-Chip (PSoC). The intrinsic parallelism of these devices makes it possible to execute several data-independent network operations in parallel. However, the approaches to designing the respective systems vary significantly with the experience and background of the engineer in charge. In this paper, we analyze and compare the pros and cons of using an embedded processor, high-level synthesis methods, and register-transfer low-level design in terms of design effort, performance, and power consumption for implementing a parallel algorithm to find the two smallest values in a dataset. This problem is easy to formulate, has a number of practical applications (for instance, in low-density parity check decoders), and is very well suited to parallel implementation based on comparator networks.

Silicon-based optoelectronics for general-purpose matrix computation: a review

Conventional electronic processors, which are the mainstream and almost invincible hardware for computation, are approaching their limits in both computational power and energy efficiency, especially in large-scale matrix computation. By combining electronic, photonic, and optoelectronic devices and circuits together, silicon-based optoelectronic matrix computation has been demonstrating great capabilities and feasibilities. Matrix computation is one of the few general-purpose computations that have the potential to exceed the computation performance of digital logic circuits in energy efficiency, computational power, and latency. Moreover, electronic processors also suffer from the tremendous energy consumption of the digital transceiver circuits during high-capacity data interconnections. We review the recent progress in photonic matrix computation, including matrix-vector multiplication, convolution, and multiply–accumulate operations in artificial neural networks, quantum information processing, combinatorial optimization, and compressed sensing, with particular attention paid to energy consumption. We also summarize the advantages of silicon-based optoelectronic matrix computation in data interconnections and photonic-electronic integration over conventional optical computing processors. Looking toward the future of silicon-based optoelectronic matrix computations, we believe that silicon-based optoelectronics is a promising and comprehensive platform for disruptively improving general-purpose matrix computation performance in the post-Moore’s law era.

How AI-enabled SDN technologies improve the security and functionality of industrial IoT network: Architectures, enabling technologies, and opportunities

The ongoing expansion of the Industrial Internet of Things (IIoT) is enabling the possibility of effective Industry 4.0, where massive sensing devices in heterogeneous environments are connected through dedicated communication protocols. This brings forth new methods and models to fuse the information yielded by the various industrial plant elements and generates emerging security challenges that we have to face, providing ad-hoc functions for scheduling and guaranteeing the network operations. Recently, the large development of Software-Defined Networking (SDN) and Artificial Intelligence (AI) technologies have made feasible the design and control of scalable and secure IIoT networks. This paper studies how AI and SDN technologies combined can be leveraged towards improving the security and functionality of these IIoT networks. After surveying the state-of-the-art research efforts in the subject, the paper introduces a candidate architecture for AI-enabled Software-Defined IIoT Network (AI-SDIN) that divides the traditional industrial networks into three functional layers. And with this aim in mind, key technologies (Blockchain-based Data Sharing, Intelligent Wireless Data Sensing, Edge Intelligence, Time-Sensitive Networks, Integrating SDN&TSN, Distributed AI) and improve applications based on AI-SDIN are also discussed. Further, the paper also highlights new opportunities and potential research challenges in control and automation of IIoT networks.

Distributed collaborative navigation based on rank-defect free network

To improve the navigation performance for network operations, a 3-dimension distributed collaborative navigation method based on rank-defect free network is proposed. Only the nodes’ rough position estimations and their relative ranges are required and the information is shared through wireless communication. The situation is highly addressed where none or only one anchor is deployed. Due to the lack of necessary measurements to fully determine the absolute positions of all agents, the algorithm providing additional constraints is developed to derive the unique solution. The characteristics of the algorithm and the factors that influence the navigation accuracy are discussed. Comprehensive simulations are conducted with different network topologies, number of agents, initial position errors and range measurement uncertainties to verify the effectiveness of the proposed method. The influences of various configurations imposed on the accuracy of the collaborative navigation are researched and quantified in the perspective of the network rank-defect.

Public transport for smart cities: Recent innovations and future challenges

The idea of a smart city is one that utilises Internet-of-Things (IoT) technologies and data analytics to optimise the efficiency of city operations and services, so as to provide a high quality of life for its citizens. Due to reduced public funding, many public transport systems are already facing challenges to maintain their services. For a smart city, the goal of public transport is not simply the movement of people, but the provision and enhancement of mobility for living. This will be particularly challenging due to changes in habitation trends and work patterns. For example, the growth of mega-cities has led to extreme traffic congestion in city centres and urban sprawl on their outskirts. In order to provide sufficient coverage and frequency of service, an integrated co-ordinated multi-modal public transport system is needed, leading to substantial increase in operational complexity. Environmental concerns and the recent pandemic may also change work and commuting patterns in the future, with more people working from home and companies adopting flexible work shifts. For smart cities, public transport must offer ubiquitous access, real-time response to demand, convenience and quality service, and energy-efficient operations. This paper will discuss the challenges in network design, operations planning, scheduling and management of smart public transport systems. A brief survey of recent research and innovations will also be presented.

A survey of deep reinforcement learning application in 5G and beyond network slicing and virtualization

The 5th Generation (5G) and beyond networks are expected to offer huge throughputs, connect large number of devices, support low latency and large numbers of business services. To realize this vision, there is a need for a paradigm shift in the way cellular networks are designed, built, and maintained. Network slicing divides the physical network infrastructure into multiple virtual networks to support diverse business services, enterprise applications and use cases. Multiple services and use cases with varying architectures and quality of service requirements on such shared infrastructure complicates the network environment. Moreover, the dynamic and heterogeneous nature of 5G and beyond networks will exacerbate network management and operations complexity. Inspired by the successful application of machine learning tools in solving complex mobile network decision making problems, deep reinforcement learning (Deep RL) methods provide potential solutions to address slice lifecycle management and operation challenges in 5G and beyond networks. This paper aims to bridge the gap between Deep RL and the 5G network slicing research, by presenting a comprehensive survey of their existing research association. First, the basic concepts of Deep RL framework are presented. 5G network slicing and virtualization principles are then discussed. Thirdly, we review challenges in 5G network slicing and the current research efforts to incorporate Deep RL in addressing them. Lastly, we present open research problems and directions for future research.

Spoofing Attack Mitigation in Address Resolution Protocol (ARP) and DDoS in Software-Defined Networking

Software Defined Networking (SDN) shows network operations to be performed for efficient network operations. Due to the increase in network devices, the percentage of attacks is also increased, and it is challenging to provide defense against such attacks. In SDN, the control plan is separated from the data plane. The control plan is implemented using some central devices called SDN controllers. In SDN Address Resolution Protocol (ARP), spoofing and Distributed Denial of Services (DDoS) attacks are carried out on an enormous scale. These are commonly launched attacks in SDN. Due to these attacks, the network performance is down, and network services are dead. This paper proposed a new auto detection methodology to detect ARP and DDoS attacks and mitigate SDN networks from these attacks. Additionally, we implemented two algorithms: one for flow rules and the second for attack detection. An individual server was installed to check the malicious traffic installation. We present the new forward flooding rules to detect and mitigate attacks. The experiments are performed using LINUX-based network implementation. Our proposal successfully improves network security and enhances network efficiency.

Smart contract-based caching and data transaction optimization in mobile edge computing

The emergence of mobile edge computing can provide technical support for the development of Internet services and for coping with massive data traffic, thereby reducing network latency and ensuring efficient network operations and service delivery. However, the mobile edge computing environment is prone to data loss and privacy leakage, and data security and reliability cannot be guaranteed. The application of blockchain technology ensures the stability and dependability of data caching and transactions. In order to guarantee the security of data caching in the mobile edge computing environment and minimize the response time of caching servers, this paper proposes a decentralized data caching strategy in the mobile edge computing environment. The strategy uses a greedy algorithm to make the transmission delay of the requested content as small as possible under the multiple constraints of the storage space of each server and whether the content is cached on the server. To address the problem that data trading platforms are unable to afford the computation and storage of massive amounts of data and cannot guarantee that the data will not be leaked, this paper proposes a secure decentralized data transaction program. The scheme promotes the increase of transaction turnover rate and improves the revenue of both participating parties by establishing a buyer–seller matching algorithm. According to the experimental results, the proposed data caching strategy can improve the cache hit rate and reduce the transmission delay; the proposed data transaction solution can increase the revenue of both the data holder and the data buyer.