Artificial Neural Network Scheme Research Articles

A highly accurate artificial neural networks scheme for solving higher multi‐order fractal‐fractional differential equations based on generalized Caputo derivative

AbstractArtificial neural networks have great potential for learning and stability in the face of tiny input data changes. As a result, artificial intelligence techniques and modeling tools have a growing variety of applications. To estimate a solution for fractal‐fractional differential equations (FFDEs) of high‐order linear (HOL) with variable coefficients, an iterative methodology based on a mix of a power series method and a neural network approach was applied in this study. In the algorithm's equation, an appropriate truncated series of the solution functions was replaced. To tackle the issue, this study uses a series expansion of an unidentified function, where this function is approximated using a neural architecture. Some examples were presented to illustrate the efficiency and usefulness of this technique to prove the concept's applicability. The proposed methodology was found to be very accurate when compared to other available traditional procedures. To determine the approximate solution to FFDEs‐HOL, the suggested technique is simple, highly efficient, and resilient.

GS-16QAM OFDM with ANN scheme combined with LFM signal for joint communication and radar sensing system.

In this Letter, an integrated waveform geometrically shaped (GS) 16 quadrature amplitude modulation (QAM) based orthogonal frequency division multiplexing (OFDM) communication signal combined with a linear frequency modulation (LFM) radar signal is proposed and experimentally demonstrated in a W-band communication and radar detection system. The proposed method can generate communication and radar signals simultaneously. The inherent error propagation and the interference of the radar signal limit the transmission performance of the joint communication and radar sensing system. Thus, an artificial neural network (ANN) scheme is proposed for the GS-16QAM OFDM signal. After 0.8-m wireless transmission, the experimental results show that the receiver sensitivity and normalized general mutual information (NGMI) of the GS-16QAM OFDM are improved compared with that of the OFDM with uniform 16QAM at the forward error correction (FEC) threshold of 3.8 × 10-3. In addition, multi-target radar detection in centimeter level radar ranging is realized.

Artificial neural network scheme to solve the hepatitis B virus model

This article aims to describe the simulation studies of the hepatitis B virus non-linear system using supervised neural networks procedures supported by Levenberg-Marquardt back propagation methodology. The proposed strategy has five distinct quantities: susceptible X(t), symptomatic infections Y(t), chronic infections W(t), recovered population R(t), and a population that has received vaccinations Z(t). The reference data set for all three distinct cases has been obtained utilizing the ND-Solver and Adams method in Mathematica software. The outcomes have been validated with performance plots for all cases. To check the accuracy and effectiveness of proposed methodology mean square error has are presented. State transition, and regression plots are illustrated to elaborated the testing, training, and validation methodology. Additionally, absolute errors for different components of hepatitis B virus model are demonstrated to depict the error occurring during distinct cases. Whereas the data assigned to training is 81%, and 9% for each testing and validation. The mean square error for all three cases is 10−12 this show the accuracy and correctness of proposed methodology.

Analysis of a bioconvection flow of magnetocross nanofluid containing gyrotactic microorganisms with activation energy using an artificial neural network scheme

The current study proposes the technique of intelligent numerical computing using back-propagated neural networks (BNNs) for the numerical treatment of a magnet cross-bioconvection flow analysis of nanofluid modal (MHDBC-NFM) including gyrotactic microorganisms with activation energy. Utilizing efficient information, the original system model MHDBC-NFM in partial differential equations (PDEs) is remade into non-linear ordinary differential equations (ODEs). Using the computational capability of the Lobatto IIIA technique, these generated non-linear ODEs are then bettered to provide a dataset of Levenberg Marquardt algorithm-based backpropagated neural networks (LMA-BNNs) for twelve scenarios of this proposed model. The validation, training, and testing processes are carried out simultaneously to acclimate the neural networks. Levenberg-Marquardt backpropagation is used to lower the mean square error (MSE) function, which is then used to derive the estimated solution for MHDBC-NFM for various scenarios. The comparative investigations and performance assessments using correlation, error histograms, MSE, and regression results illustrate the efficacy of the suggested BNNs methodology. The observed diminishing behavior of the nondimensional gyrotactic microorganism field is correlated with the bioconvection Lewis number. Additionally, the results show that temperature profiles become more intense for higher radiation parameter values.

Detection of Management-Frames-Based Denial-of-Service Attack in Wireless LAN Network Using Artificial Neural Network

Wireless Local Area Networks (WLANs) have become an increasingly popular mode of communication and networking, with a wide range of applications in various fields. However, the increasing popularity of WLANs has also led to an increase in security threats, including denial of service (DoS) attacks. In this study, management-frames-based DoS attacks, in which the attacker floods the network with management frames, are particularly concerning as they can cause widespread disruptions in the network. Attacks known as denial of service (DoS) can target wireless LANs. None of the wireless security mechanisms in use today contemplate defence against them. At the MAC layer, there are multiple vulnerabilities that can be exploited to launch DoS attacks. This paper focuses on designing and developing an artificial neural network (NN) scheme for detecting management-frames-based DoS attacks. The proposed scheme aims to effectively detect fake de-authentication/disassociation frames and improve network performance by avoiding communication interruption caused by such attacks. The proposed NN scheme leverages machine learning techniques to analyse patterns and features in the management frames exchanged between wireless devices. By training the NN, the system can learn to accurately detect potential DoS attacks. This approach offers a more sophisticated and effective solution to the problem of DoS attacks in wireless LANs and has the potential to significantly enhance the security and reliability of these networks. According to the experimental results, the proposed technique exhibits higher effectiveness in detection compared to existing methods, as evidenced by a significantly increased true positive rate and a decreased false positive rate.

An Uncertain Mean-AVaR Portfolio Selection via an Artificial Neural Network Scheme

This paper focuses on the computation issue of portfolio optimization with scenario-based mean-average value at Risk (AVaR) in uncertain environment. The portfolio optimization problem is designed in two cases: risk-taker and risk-averse models. The main idea is to replace the portfolio selection models with linear programming (LP) problems. Since the computing time required for solving LP greatly depends on the dimension and the structure of the problem, the conventional numerical methods are usually less effective in real-time applications. One promising approach to handle online applications is to employ recurrent neural networks based on circuit implementation. Hence, according to the convex optimization theory and some concepts of ordinary differential equations, a neural network model for solving the LP problems related to portfolio selection problems is presented. The equilibrium point of the proposed model is proved to be equivalent to the optimal solution of the original problem. It is also shown that the proposed neural network model is stable in the sense of Lyapunov and it is globally convergent to an exact optimal solution of the portfolio selection problem with uncertain returns. Some illustrative examples are provided to show the feasibility and the efficiency of the proposed method in this paper.

Mean-AVaR in credibilistic portfolio management via an artificial neural network scheme

ABSTRACT This paper focuses on the computation issue of portfolio optimisation with scenario-based mean-Average Value at Risk (AVaR) in credibilistic environment. The portfolio optimisation problem is designed in two cases: risk taker model and risk-averse model. The main idea is to replace the portfolio selection models with linear programming (LP) problems. Since the computing time required for solving LP greatly depends on the dimension and the structure of the problem, the conventional numerical methods are usually less effective in real-time applications. One promising approach to handle online applications is to employ recurrent neural networks based on circuit implementation. Hence, according to the convex optimisation theory and some concepts of ordinary differential equations, a neural network model for solving the LP problems related to portfolio selection problems is presented. The equilibrium point of the proposed model is proved to be equivalent to the optimal solution of the original problem. It is also shown that the proposed neural network model is stable in the sense of Lyapunov and it is globally convergent to an exact optimal solution of the portfolio selection problem with fuzzy returns. Some illustrative examples are provided to show the feasibility and the efficiency of the proposed method in this paper.

Privacy-Preserving Outsourced Artificial Neural Network Training for Secure Image Classification

Artificial neural network (ANN) is powerful in the artificial intelligence field and has been successfully applied to interpret complex image data in the real world. Since the majority of images are commonly known as private with the information intended to be used by the owner, such as handwritten characters and face, the private constraints form a major obstacle in developing high-precision image classifiers which require access to a large amount of image data belonging to multiple users. State-of-the-art privacy-preserving ANN schemes often use full homomorphic encryption which result in a substantial overhead of computation and data traffic for the data owners, and are restricted to approximation models by low-degree polynomials which lead to a large accuracy loss of the trained model compared to the original ANN model in the plain domain. Consequently, it is still a huge challenge to train an ANN model in the encrypted-domain. To mitigate this problem, we propose a privacy-preserving ANN system for secure constructing image classifiers, named IPPNN, where the server is able to train an ANN-based classifier on the combined image data of all data owners without being able to observe any images using primitives, such as randomization and functional encryption. Our system achieves faster training time and supports lossless training. Moreover, IPPNN removes the need for multiple communications among data owners and servers. We analyze the security of the protocol and perform experiments on a large scale image recognition task. The results show that the IPPNN is feasible to use in practice while achieving high accuracy.

Improving Probabilistic Quantitative Precipitation Forecasts Using Short Training Data through Artificial Neural Networks

Abstract Conventional statistical postprocessing techniques offer limited ability to improve the skills of probabilistic guidance for heavy precipitation. This paper introduces two artificial neural network (ANN)-based, geographically aware, and computationally efficient postprocessing schemes, namely, the ANN-multiclass (ANN-Mclass) and the ANN–censored, shifted gamma distribution (ANN-CSGD). Both schemes are implemented to postprocess Global Ensemble Forecast System (GEFS) forecasts to produce probabilistic quantitative precipitation forecasts (PQPFs) over the contiguous United States (CONUS) using a short (60 days), rolling training window. The performances of these schemes are assessed through a set of hindcast experiments, wherein postprocessed 24-h PQPFs from the two ANN schemes were compared against those produced using the benchmark quantile mapping algorithm for lead times ranging from 1 to 8 days. Outcomes of the hindcast experiments show that ANN schemes overall outperform the benchmark as well as the raw forecast over the CONUS in predicting probability of precipitation over a range of thresholds. The relative performance varies among geographic regions, with the two ANN schemes broadly improving upon quantile mapping over the central, south, and southeast, and slightly underperforming along the Pacific coast where skills of raw forecasts are the highest. Between the two schemes, the hybrid ANN-CSGD outperforms at higher rainfall thresholds (i.e., >50 mm day−1), though the outperformance comes at a slight expense of sharpness and spatial specificity. Collectively, these results confirm the ability of the ANN algorithms to produce skillful PQPFs with a limited training window and point to the prowess of the hybrid scheme for calibrating PQPFs for rare-to-extreme rainfall events.

Protection for 330 kV transmission line and recommendation for Nigerian transmission system: a review

The transmission line is an integral part of the electrical power system; however, a fault has a negative impact on the system, like blackout, power loss, financial losses, and socio-economic impact. This fault occurs due to ageing conductors, lightning stroke, switching surge and human interference. We reviewed the protection scheme implemented in the Nigerian transmission network, which has challenges relating to the environment's terrain and a long-distance transmission line of about 20,000 km. The different approach of fault classification, detection and location was analysed and critically summarised. This review paper proposes a hybrid Artificial Neural Network and distance protection scheme that can automatically identify, locate, isolate, predict, correct faults, and real-time monitor and control the entire network. It can also detect the shortest possible trip time of 0.02 s and 0.03 s of line current and fault losses, respectively, during fault to avert damage on the line. However, this method has its challenges, such as the volume of data generated from load flow analysis, training time, and the total distance covered by the network. However, these can be averted by segmenting the entire network for easy evaluation and monitoring to achieve set goals.

Traffic Classification in Underwater Networks Using SDN and Data-Driven Hybrid Metaheuristics

Software-Defined Networks ( SDNs ), with their segregated data and control planes, has proved to be capable of managing massive amounts of data by leveraging distributed information available across the network for informed decision-making at the network controller. However, with the proliferation of next-generation, real-time Internet of Things ( IoT ) applications that vary greatly in terms of data frequency and volumes, data traffic classification can substantially assist SDN controllers toward efficient routing and traffic engineering decisions. Existing works on network classification are limited by their application-centric nature, thus overlooking the key criterion for real-time IoT applications, namely, Quality of Service ( QoS ). In this article, we focus on augmenting SDN controllers’ decision-making capacity and Underwater Sensor Networks with machine learning algorithms to achieve real-time, QoS-aware, network traffic classification. Three classifiers, namely, Feed-forward Neural Network, Naïve Bayes, and Logistics Regression have been employed with a novel Artificial Neural Network and Particle Swarm Optimization hybridization scheme by carrying first- and second-order stability analysis for performance improvement of these classifiers. In short, the proposed framework exploits optimization algorithms and semi-supervised machine learning ( ML ) for precise traffic classification while keeping communication overhead between controller and switches minimal. Results obtained from real-life datasets demonstrate the efficacy of our proposed scheme.

An advance artificial neural network scheme to examine the waste plastic management in the ocean

In this study, an advanced computational artificial neural network (ANN) procedure is designed using the novel characteristics of the Levenberg–Marquardt backpropagation (LBMBP), i.e., ANN-LBMBP, for solving the waste plastic management in the ocean system that plays an important role in the economy of any country. The nonlinear mathematical form of the waste plastic management in the ocean system is categorized into three groups: waste plastic material W(χ), marine debris M(χ), and reprocess or recycle R(χ). The learning based on the stochastic ANN-LBMBP procedures for solving mathematical waste plastic management in the ocean is used to authenticate the sample statics, testing, certification, and training. Three different statistics for the model are considered as training 70%, while for both validation and testing are 15%. To observe the performances of the mathematical model, a reference dataset using the Adams method is designed. To reduce the mean square error (MSE) values, the numerical performances through the ANN-LBMBP procedures are obtained. The accuracy of the designed ANN-LBMBP procedures is observed using the absolute error. The capability, precision, steadfastness, and aptitude of the ANN-LBMBP procedures are accomplished based on the multiple topographies of the correlation and MSE.

Artificial neural network scheme to solve the nonlinear influenza disease model

The aim of this study is to present the numerical simulations of the influenza disease nonlinear system (IDNS) using the stochastic artificial neural networks (ANNs) procedures supported with Levenberg-Marquardt backpropagation (LMB), i.e., ANNs-LMB. The IDNS is constructed with four classes, susceptible S(t), infected I(t), recovered R(t) and cross-immune people C(t), based stiff nonlinear ordinary differential system. The numerical computations have been performed through the stochastic ANNs-LMB for solving six different variations of the IDNS. The obtained numerical solutions through the stochastic ANNs-LMB for solving the IDNS have been presented using the training, verification and testing measures to reduce mean square error (MSE) from data-based reference solutions. To observed the correctness, efficiency, competence and proficiency of the designed computing paradigm ANNs-LMB, an exhaustive analysis is presented using the correlation studies, error histograms (EHs), mean squared error (MSE), regression and state transitions (STs) information. The worth and significance of ANNs-LMB is substantiated through comparisons of the outcomes admitted the good agreement from data derived results with 5–7 decimal places of accuracy for each scenario of IDNS.

A novel weather parameters prediction scheme and their effects on crops

Weather forecast is significantly imperative in today’s smart technological world. A precise forecast model entails a plentiful data in order to attain the most accurate predictions. However, a forecast of future rainfall from historical data samples has always been challenging and key area of research. Hence, in modern weather forecasting a combo of computer models, observation, and knowledge of trends and patterns are introduced. This research work has presented a fitness function based adaptive artificial neural network scheme in order to forecast rainfall and temperature for upcoming decade (2021-2030) using historical weather data of 20 different districts of Karnataka state. Furthermore, effects of these forecasted weather parameters are realized over five major crops of Karnataka namely rice, wheat, jowar, maize, and ragi with the intention of evaluation for efficient crop management in terms of the passing relevant messages to the farmers and alternate measures such as suggesting other geographical locations to grow the same crop or growing other suitable crops at same geographical location. A graphical user interface (GUI) application has been developed for the proposed work in order to ease out the flow of work.

Screening of important parameters in optimal design of compressed air energy storage system using an ensemble learning method

Accurate prediction of thermophysical properties of compressed air is specifically crucial in optimal design and analyzing performance of a Compressed Air Energy Storage (CAES) system. In this regard, an advanced machine learning method, namely stochastic gradient boosting (SGBTree) technique was examined based on reported data points in the literature for estimation of specific heat ratio of air, water content of atmospheric air, water content of compressed saturated air, viscosity of compressed air as well as thermal conductivity of compressed air. The used ensemble learning method precisely functions for a wide range of temperature, pressure and relative humidity. The SGBTree models estimate the aforementioned target values well, matching to those available in the literature as the correlation coefficient (R) and Mean Relative Absolute Error (MRAE) values are greater than 0.999 and less than 0.007, respectively. Moreover, a comparison was carried out between the outputs of the SGBTree model and two intelligent models namely Artificial Neural Network (ANN) and Least Squares Support Vector Machine (LSSVM) schemes as well as a reported simple empirical correlation for estimating specific heat ratio of air. In terms of accuracy, it is found that the SGBTree model is better than other methods for estimation of specific heat ratio of air.

PU-DetNet: Deep Unfolding Aided Smart Sensing Framework for Cognitive Radio

Spectrum sensing in cognitive radio (CR) paradigm can be broadly categorized as analytical-based and data-driven approaches. The former is sensitive to model inaccuracies in evolving network environment, while the latter (machine learning (ML)/deep learning (DL) based approach) suffers from high computational cost. For devices with low computational abilities, such approaches could be rendered less useful. In this context, we propose a deep unfolding architecture namely the Primary User-Detection Network (PU-DetNet) that harvests the strength of both: analytical and data-driven approaches. In particular, a technique is described that reduces computation in terms of inference time and the number of floating point operations (FLOPs). It involves binding the loss function such that each layer of the proposed architecture possesses its own loss function whose aggregate is optimized during training. Compared to the state-of-the-art, experimental results demonstrate that at SNR= -10 dB, the probability of detection is significantly improved as compared to the long short term memory (LSTM) scheme (between 39% and 56%), convolutional neural network (CNN) scheme (between 45% and 84%), and artificial neural network (ANN) scheme (between 53% and 128%) over empirical, 5G new radio, DeepSig, satellite communications, and radar datasets. The accuracy of proposed scheme also outperforms other existing schemes in terms of the F1-score. Additionally, inference time reduces by 91.69%, 90.90%, and 93.15%, while FLOPs reduces by 62.50%, 56.25%, 64.70% w.r.t. LSTM, CNN and ANN schemes, respectively. Moreover, the proposed scheme also shows improvement in throughput by 56.39%, 51.23%, and 69.52% as compared to LSTM, CNN and ANN schemes respectively, at SNR = -6 dB.

An artificial neural network prediction on physical, mechanical, and thermal characteristics of giant reed fiber reinforced polyethylene terephthalate composite

Plant fiber reinforced hybrid polymer composites have had broad applications recently because of their lower cost advantages, lower weight, and biodegradable nature. The present work studies the influence of reinforcing giant reed fiber concentration in polyethylene terephthalate (PET) polymer for their physical, mechanical, and thermal characteristics and determines the optimum loading of giant reed fiber using an artificial neural network (ANN) scheme. Giant reed fiber reinforced PET matrix laminates were manufactured from compression molding with different fiber loadings such as 5 wt.%, 10 wt.%, and 20 wt.%. The mechanical characteristics such as tensile and flexural strength and the laminate’s tensile and flexural modulus were appraised and examined. The maximum value of tensile strength, flexural strength, tensile modulus, and flexural modulus were 5.4 MPa, 26 MPa, 8343 MPa, and 6300 MPa, respectively, for PET2 (10 wt.% of giant reed fiber in PET polymer) composite. Fiber pullout, gaps, and fracture behavior were examined from a scanning electron microscope in the microstructural analysis. A machine learning technique has been recommended to combine artificial intelligence while designing giant reed fiber reinforced polymeric laminates. Using the suggested method, an ANN model has been generated to attain the targeted giant reed fiber concentration for PET composite while gratifying the necessary targeted characteristics. The developed method is very effective and decreases the effort and time of material characterization for huge specimens. It will support the researchers in designing their forthcoming test efficiently.

Rüzgâr Enerjisi Dönüşüm Sistemlerinde bir DC-DC Güç Dönüştürücüsünün Denetimli Öğrenme Destekli Kontrolü

Over the last decades, to adopt high penetration of renewable energy sources (RESs) in electrical energy systems, distributed energy resources (DERs) have become prominent. Due to easy attainability status of small wind turbines (WTs), wind energy conversion systems (WECSs) are feasible applications for small customers, especially in windy areas. The next decade is likely to witness a considerable rise in DERs. In this context, WECSs are preferred broadly, thus harvesting wind energy into electrical energy effectively is a substantial issue. WTs can be got involved in the grid-connected or autonomous mode with a variety of topologies. In this paper, we examine to control of DC-DC boost converter of a WECS with the help of artificial intelligence (AI)-aided PI controller based on supervised learning method. Regarding the proposed method, artificial neural networks (ANNs) as a subset of AI are utilized. To test and ensure the applicability of the proposed control method, a small WECS with a permanent magnet synchronous generator (PMSG) connected a DC bus was implemented in MATLAB/Simulink environment. The proposed ANN scheme has reached a high accuracy rate with an overall mean squared error (MSE) equal to 7.4e-08. The results present that dynamic response and less complexity with a high accuracy rate have been obtained under study. The main target of this study is to reduce the number of sensors in the control layer. Thus, a cost-effective and more reliable structure is obtained with fewer sensor requirements.

Use of machine learning for unraveling hidden correlations between particle size distributions and the mechanical behavior of granular materials

A data-driven framework was used to predict the macroscopic mechanical behavior of dense packings of polydisperse granular materials. The discrete element method, DEM, was used to generate 92,378 sphere packings that covered many different kinds of particle size distributions, PSD, lying within 2 particle sizes. These packings were subjected to triaxial compression and the corresponding stress–strain curves were fitted to Duncan–Chang hyperbolic models. An artificial neural network (NN) scheme was able to anticipate the value of the model parameters for all these PSDs, with an accuracy similar to the precision of the experiment and even when the NN was trained with a few hundred DEM simulations. The estimations were indeed more accurate than those given by multiple linear regressions (MLR) between the model parameters and common geotechnical and statistical descriptors derived from the PSD. This was achieved in spite of the presence of noise in the training data. Although the results of this massive simulation are limited to specific systems, ways of packing and testing conditions, the NN revealed the existence of hidden correlations between PSD of the macroscopic mechanical behavior.

External aerodynamic investigation over Ahmed body for optimal topology selection between upper and under bodywork using ANN approach

Aerodynamic improvements primarily result in decreased fuel usage and carbon dioxide emissions into the atmosphere. Numerous governments support ongoing aerodynamics development initiatives as a means of addressing the energy problem and reducing air pollution, Ahmed body investigation helps research to investigate versatile approaches and flexibility of design. This study is carried over a generic design of Ahmed body model. We attempted a passive arrangement system to reduce drag coefficient with a correlation of cases such as in primary objective varying parameter of slant angle from 20∘ to 30∘ proposed to monitor the behavior of drag coefficient. Once we finalized the optimum slant angle, which gives a lower drag coefficient, the next proposed configuration is to vary passive arrangement between lower and upper blend length to see the deflection of the boundary layer in correlation with the drag coefficient. The final topology is selected, which gives the lowest drag coefficient. The post-process correlation study was proposed by using an artificial neural network (ANN) scheme. The ANN model is developed between an achieved set of data from CFD investigation, ANN model indicates a strong correlation between the varying percentage of blend angle and increment percentage of the drag coefficient.