Computer Storage Requirements Research Articles

Deep Reinforcement Learning-based Mining Task Offloading Scheme for Intelligent Connected Vehicles in UAV-aided MEC

The convergence of unmanned aerial vehicle (UAV)-aided mobile edge computing (MEC) networks and blockchain transforms the existing mobile networking paradigm. However, in the temporary hotspot scenario for intelligent connected vehicles (ICVs) in UAV-aided MEC networks, deploying blockchain-based services and applications in vehicles is generally impossible due to its high computational resource and storage requirements. One possible solution is to offload part of all the computational tasks to MEC servers wherever possible. Unfortunately, due to the limited availability and high mobility of the vehicles, there is still lacking simple solutions that can support low-latency and higher reliability networking services for ICVs. In this article, we study the task offloading problem of minimizing the total system latency and the optimal task offloading scheme, subject to constraints on the hover position coordinates of the UAV, the fixed bonuses, flexible transaction fees, transaction rates, mining difficulty, costs and battery energy consumption of the UAV. The problem is confirmed to be a challenging linear integer planning problem, we formulate the problem as a constrained Markov decision process. Deep Reinforcement Learning (DRL) has excellently solved sequential decision-making problems in dynamic ICVs environment, therefore, we propose a novel distributed DRL-based P-D3QN approach by using Prioritized Experience Replay strategy and the dueling double deep Q-network (D3QN) algorithm to solve the optimal task offloading policy effectively. Finally, experiment results show that compared with the benchmark scheme, the P-D3QN algorithm can bring about 26.24% latency improvement and increase about 42.26% offloading utility.

Beaf:BD – A Blockchain Enabled Authentication Framework for Big Data

The widespread utilization of Internet-based applications in our daily routines has resulted in enormous amounts of data being generated every minute. This data is not only produced by humans but also by various machines such as sensors, satellites, CCTV, etc. For many organizations, Apache Hadoop is the solution for handling big data. Big data refers to the extensive set of dissimilar data that can be processed to derive meaningful insights. For its security needs, Hadoop relies on trusted third-party security providers such as Kerberos. Kerberos has several security vulnerabilities. The focus of this paper is to eliminate security issues, particularly dictionary attacks and single points of failure, by proposing a model based on blockchain technology and threshold cryptography.In comparison to other existing schemes, the proposed approach offers superior computational overhead and storage requirements while maintaining the system's security level.

SuperAdjoint: Super-resolution neural networks in adjoint-based error estimation

Numerical simulations and optimisation methods, such as mesh adaptation, rely on the accurate and inexpensive use of error estimation methods. Adjoint-based error estimation is the most accurate method, and generally the most costly. A strong contributor to this cost is the need to compute a higher resolution adjoint solution. Here, it is proposed to use super-resolution neural networks to super-resolve a fine adjoint solution from a lower-cost coarse adjoint solution: a superAdjoint. The method is compared to reference error estimators on an unsteady Burgers’ equation using the method of manufactured solutions. Two forms of the superAdjoint were implemented, a twice and a four times refining super-resolution neural network. These were used to demonstrate both the computational cost reduction and the potential for the reduction of the storage footprint of the primal problem. The first, referred to as 2×CNN, was able to reconstruct the spatially enriched adjoint solution, thus providing a robust and inexpensive local output error. The second, the 4×CNN, was able to demonstrate the reconstruction ability of super-resolution neural networks for higher upscaling factors. This was leveraged in order to subsample the primal solution in space, thus reducing substantially the storage footprint of the discrete primal solution. Both superAdjoints could achieve the desired level of accuracy when compared to the reconstruction of the refined adjoint-based error estimate. Moreover, superAdjoint was shown to be able to generalise to a new QoI that was untrained for. This gives great confidence in the use of super-resolution neural networks for the reduction of both computational cost and storage requirements of adjoint-based error estimation, and goal-oriented mesh adaptation.

Fourier-Gegenbauer pseudospectral method for solving time-dependent one-dimensional fractional partial differential equations with variable coefficients and periodic solutions

In this paper, we present a novel pseudospectral (PS) method for solving a new class of initial-value problems (IVPs) of time-dependent one-dimensional fractional partial differential equations (FPDEs) with variable coefficients and periodic solutions. A main ingredient of our work is the use of the recently developed periodic RL/Caputo fractional derivative (FD) operators with sliding positive fixed memory length of Bourafaet al. (2021) or their reduced forms obtained by Elgindy (2023) as the natural FD operators to accurately model FPDEs with periodic solutions. The proposed method converts the IVP into a well-conditioned linear system of equations using the PS method based on Fourier collocations and Gegenbauer quadratures. The reduced linear system has a simple special structure and can be solved accurately and rapidly by using standard linear system solvers. A rigorous study of the computational storage requirements as well as the error and convergence of the proposed method is presented. The idea and results presented in this paper are expected to be useful in the future to address more general problems involving FPDEs with periodic solutions.

Computationally efficient approach for risk-informed decision making

Probabilistic risk assessment (PRA) is used as an essential tool for risk-informed decision-making in the nuclear industry. The fault and event trees play a crucial role in PRA to estimate the probability of system failure based on the failure probabilities of components. The fault trees or event trees for an actual power plant unit can be fairly large in size with several different types of logic gates, interconnected events, dependent events, etc. A large fault tree can include hundreds of gates, basic events (BEs), multiple occurring events (MOEs), and dependent events. Complex connectivities can give rise to excessive computational demand and storage requirements for the analysis. Fault and event trees can be solved using the minimal cut-set approaches, or advanced quantification techniques such as Binary decision diagrams or Bayesian networks. However, these techniques can be computationally inefficient for larger fault trees and can run out of memory/storage space. This study focuses on developing and proposing a new approach for accurate estimation of the system-level risk while improving the computational efficiency significantly. More specifically, an attempt is made to reduce the complexity of the analysis of MOEs and dependent events in fault trees. The proposed algorithms in this study present a significant improvement over traditional approaches which makes it highly promising for additional development. The computational efficiency of the proposed approach over the traditional approach is illustrated for fault trees with a varying number of events and different types of logic gate connections.

Filter Pruning with Convolutional Approximation Small Model Framework

Convolutional neural networks (CNNs) are extensively utilized in computer vision; however, they pose challenges in terms of computational time and storage requirements. To address this issue, one well-known approach is filter pruning. However, fine-tuning pruned models necessitates substantial computing power and a large retraining dataset. To restore model performance after pruning each layer, we propose the Convolutional Approximation Small Model (CASM) framework. CASM involves training a compact model with the remaining kernels and optimizing their weights to restore feature maps that resemble the original kernels. This method requires less complexity and fewer training samples compared to basic fine-tuning. We evaluate the performance of CASM on the CIFAR-10 and ImageNet datasets using VGG-16 and ResNet-50 models. The experimental results demonstrate that CASM surpasses the basic fine-tuning framework in terms of time acceleration (3.3× faster), requiring a smaller dataset for performance recovery after pruning, and achieving enhanced accuracy.

Opt-RNN-DBFSVM: Optimal recurrent neural network density based fuzzy support vector machine

Two major problems are encountered when using fuzzy SVM: (a) the number of local minima increases exponentially with the number of samples and (b) the quantity of required computer storage, required for a regular quadratic programming solver, increases by an exponential magnitude as the problem size expands. The Kernel-Adatron family of algorithms gaining attention lately which has allowed to handle very large classification and regression problems. However, these methods treat different types of samples (Noise, border, and core) with the same manner, which causes searches in unpromising areas and increases the number of iterations. In this work, we introduce a hybrid method to overcome these shortcoming, namely Optimal Recurrent Neural Network Density Based fuzzy Support Vector Machine (Opt-RNN-DBFSVM). This method consists of four steps: (a) characterization of different samples, (b) elimination of samples with a low probability of being a support vector, (c) construction of an appropriate recurrent neural network based on an original energy function, and (d) solution of the system of differential equations, managing the dynamics of the RNN, using the Euler–Cauchy method involving an optimal time step. Thanks to its recurrent architecture, the RNN remembers the regions explored during the search process. We demonstrated that RNN-FSVM converges to feasible support vectors and Opt-RNN-DBFSVM has a very low time complexity compared to RNN-FSVM with constant time step, and KAs-FSVM. Several experiments were performed on academic data sets. We used several classification performance measures to compare Opt-RNN-DBFSVM to different classification methods and the results obtained show the good performance of the proposed method.

A High-Performance Automated Large-Area Land Cover Mapping Framework

Land cover mapping plays a pivotal role in global resource monitoring, sustainable development research, and effective management. However, the complexity of the mapping process, coupled with significant computational and data storage requirements, often leads to delays between data processing and product publication, thereby bringing challenges to creating multi-timesteps large-area products for monitoring dynamic land cover. Therefore, improving the efficiency of each stage in land cover mapping and automating the mapping process is currently an urgent issue to be addressed. This study proposes a high-performance automated large-area land cover mapping framework (HALF). By leveraging Docker and workflow technologies, the HALF effectively tackles model heterogeneity in complex land cover mapping processes, thereby simplifying model deployment and achieving a high degree of decoupling between production models. It optimizes key processes by incorporating high-performance computing techniques. To validate these methods, this study utilized Landsat imagery data and extracted samples using GLC_FCS and FROM_GLC, all of which were acquired at a spatial resolution of 30 m. Several 10° × 10° regions were chosen globally to illustrate the viability of generating large-area land cover using the HALF. In the sample collection phase, the HALF introduced an automated method for generating samples, which overlayed multiple prior products to generate a substantial number of samples, thus saving valuable manpower resources. Additionally, the HALF utilized high-performance computing technology to enhance the efficiency of the sample–image matching phase, thereby achieving a speed that was ten times faster than traditional matching methods. In the mapping stage, the HALF employed adaptive classification models to train the data in each region separately. Moreover, to address the challenge of handling a large number of classification results in a large area, the HALF utilized a parallel mosaicking method for classification results based on the concept of grid division, and the average processing time for a single image was approximately 6.5 s.

Block-Wise Computation of Cyclic Redundancy Code Using Factored Toeplitz Matricesin Lieu of Look-Up Table

Cyclic Redundancy Check (CRC) or Cyclic Redundancy Code is a Cyclic Error Detection Code used to preserve the integrity of data in storage and transmission applications. CRC of a stream of message bits is usually calculated block-wise in parallel with the help of a Look-Up Table (LUT) in software or State Space transformation matrix in hardware. Presented here is a novel method and architecture of parallel computation of Cyclic Redundancy Codes without any Look-Up Table for an arbitrary generating polynomial which is programmable at runtime. The method reduces computational complexity and storage requirements by implicitly factorizing the transformation matrix needed to compute the remainder into two simpler Toeplitz matrices. The resulting hardware architecture is suitable for embedded applications.

A Fast Forward and Inversion Strategy for Three-Dimensional Gravity Field

Obtaining a three-dimensional (3D) density distribution within a reasonable time is one of the most critical problems in gravity exploration. In this paper, we present an efficient 3D forward modeling and inversion method for gravity data. In forward modeling, the 3D model is discretized into multiple horizontal layers, with the gravity field at a point on the surface being the sum of the gravity fields from all layers. To calculate the gravity field from each horizontal layer, we use the fast Fourier transform (FFT) method and the Block Toeplitz with Toeplitz Blocks (BTTB) matrix, which dramatically reduces both the computation time and storage requirement. In the inversion, the observed gravity data are separated into multiple gravity components of different depths using the cutting separation method. An iterative method is used to adjust the model to fit the above gravity component for each cutting radius. The initial model is constructed from the transformation of gravity components. These methods were applied to both synthetic data and field data. The numerical simulation validated the proposed methods, and the inversion results of field data were consistent with information obtained from well logging. The computational time and memory usage were also reasonable.

Reduced basis methods with parameterized boundary conditions for room acoustics

Room acoustic simulations can be performed by means of numerical methods, which typically solve the wave equation in an enclosure using discretization techniques. These methods provide high-fidelity solvers that include all the wave phenomena but are computationally costly. This paper presents the potential of a reduced basis method for simulating room acoustics when the boundary properties are changed iteratively, e.g., design changes in rooms where different boundary properties are simulated, which increase the computational cost when using traditional high-fidelity solvers. The presented framework allows reducing the computational burden by applying reduced basis methods and solving the problem in a reduced low-dimensional subspace where the absorption properties of complex boundary conditions are parameterized, e.g., the thickness of a porous material. The potential of the proposed framework is analyzed in terms of computational efficiency, accuracy and storage requirements. We show a computational reduction of two orders of magnitude for a 2D case with an upper frequency of 2 kHz and three orders of magnitude for a 3D simple case with an upper frequency of 1 kHz showing a potential of having an increase in terms of speedup when increasing the size of the domain. The storage requirement for a 2D case with an upper frequency of 8 kHz is 10.8GB while for a 3D case with an upper frequency of 1 kHz is 8GB.

Analysis and Application of Matrix-Form Neural Networks for Fast Matrix-Variable Convex Optimization.

Matrix-variable optimization is a generalization of vector-variable optimization and has been found to have many important applications. To reduce computation time and storage requirement, this article presents two matrix-form recurrent neural networks (RNNs), one continuous-time model and another discrete-time model, for solving matrix-variable optimization problems with linear constraints. The two proposed matrix-form RNNs have low complexity and are suitable for parallel implementation in terms of matrix state space. The proposed continuous-time matrix-form RNN can significantly generalize existing continuous-time vector-form RNN. The proposed discrete-time matrix-form RNN can be effectively used in blind image restoration, where the storage requirement and computational cost are largely reduced. Theoretically, the two proposed matrix-form RNNs are guaranteed to be globally convergent to the optimal solution under mild conditions. Computed results show that the proposed matrix-form RNN-based algorithm is superior to related vector-form RNN and matrix-form RNN-based algorithms, in terms of computation time.

A 3D finite-strain beam model for thermo-mechanical deformations of 2D shape memory alloys in 3D space

In this paper, a robust constitutive model adapted to a 3D beam theory in the finite strain regime is proposed to simulate thermo-mechanical behaviors of shape memory alloys (SMAs) under triaxial normal-shear loadings. For the first time, four internal parameters are considered to capture martensitic transformation, orientation/reorientation of martensite variants, normal-shear deformation coupling, and asymmetric/anisotropic strain generation in polycrystalline SMAs. Solution algorithms of the model are presented for stress- and strain-control cases based on the elastic-predictor inelastic-corrector return mapping process. To develop the Cauchy-Green deformation tensor, a 3D exact displacement field is proposed based on the centroid movements and rotations of the cross-section. A finite element formulation along with the Newton-Raphson and Riks techniques is also established to trace the non-linear equilibrium path of 3D SMA beam structures in the large defamation regime. Higher-order shape functions are also considered to avoid shear and membrane locking issues. To explore and demonstrate the capabilities of the proposed model, the results of the triaxial and 3D SMA constitutive models are compared with existing experimental results on uniaxial tension, and combined tension-torsion tests. Afterwards, the results of 3D solid element and proposed 3D beam element are compared with experimental results of deep stretch SMA helical springs under three loading/unloading cycles. An excellent qualitative and quantitative correlation is observed between numerical and experimental results verifying the model accuracy and the solution procedure. The combination of the proposed triaxial SMA constitutive model with 3D beam theory leads to a significant reduction in computational time and storage requirements without any accuracy loss. Extra simulations are then conducted for a simplified model of a 3D biomedical vascular SMA stent under various loading/unloading conditions. The simulations reveal that the new 3D finite-strain SMA-beam element is well appropriate for modeling of 3D spatial lattice-based stents. Due to simplicity and accuracy, the 3D SMA beam model could serve in future efforts for a computationally efficient modeling of large defamations of complicated 3D SMA structures with 2D members (e.g., SMA lattices). • A robust constitutive model is developed to simulate behaviors of SMAs under triaxial normal-shears loadings. • The solution algorithms combined is extended to trace the non-linear equilibrium path of 3D SMA beam structures. • The results of present 3D Beam model has been compared with existing experimental results as well as 3D solid elements. • A good qualitative and quantitative correlation is observed between numerical and experimental results. • The proposed triaxial SMA beam model leads to a significant reduction in computational time without loss inaccuracy.

Physics-based digital twins for autonomous thermal food processing: Efficient, non-intrusive reduced-order modeling

One possible way of making thermal processing controllable is to gather real-time information on the product's current state. Often, sensory equipment cannot capture all relevant information easily or at all. Digital Twins close this gap with virtual probes in real-time simulations, synchronized with the process. This paper proposes a physics-based, data-driven Digital Twin framework for autonomous food processing. We suggest a lean Digital Twin concept that is executable at the device level, entailing minimal computational load, data storage, and sensor data requirements. This study focuses on a parsimonious experimental design for training non-intrusive reduced-order models (ROMs) of a thermal process. A correlation (R = − 0.76) between a high standard deviation of the surface temperatures in the training data and a low root mean square error in ROM testing enables efficient selection of training data. The mean test root mean square error of the best ROM is less than 1 Kelvin (0.2% mean average percentage error) on representative test sets. Simulation speed-ups of Sp ≈ 1.8 × 104 allow on-device model predictive control. Industrial relevanceThe proposed Digital Twin framework is designed to be applicable within the industry. Typically, non-intrusive reduced-order modeling is required as soon as the modeling of the process is performed in software, where root-level access to the solver is not provided, such as commercial simulation software. The data-driven training of the reduced-order model is achieved with only one data set, as correlations are utilized to predict the training success a priori.

A fast and storage-saving method for direct volumetric integration of FWH acoustic analogy

This paper proposes a new method that makes the direct volume integration of the Lighthill stress term in FWH acoustic analogy affordable in terms of both computational time and disk storage requirement. The method is based on a dual mesh concept, i.e. while a fine mesh is used in CFD, a relatively coarse mesh is used for the acoustic calculation. The terms being independent of the observer positions in the integral formula of FWH acoustic analogy are calculated on the CFD mesh and mapped onto the acoustic mesh during the simulation. These sound source terms are used for the on-the-fly acoustic computation or saved for later use in the post-processing. A formulation of the Lighthill stress tensor in cavitating flow is also derived. The method is verified with the acoustic assessment of an inclined marine propeller case. A reduction of the calculation time by 98% and disk storage by 99% is achieved while maintaining small acoustic error.

Reduced basis methods for numerical room acoustic simulations with parametrized boundaries.

The use of model-based numerical simulations of wave propagation in rooms for engineering applications requires that acoustic conditions for multiple parameters are evaluated iteratively, which is computationally expensive. We present a reduced basis method (RBM) to achieve a computational cost reduction relative to a traditional full-order model (FOM) for wave-based room acoustic simulations with parametrized boundaries. The FOM solver is based on the spectral-element method; however, other numerical methods could be applied. The RBM reduces the computational burden by solving the problem in a low-dimensional subspace for parametrized frequency-independent and frequency-dependent boundary conditions. The problem is formulated in the Laplace domain, which ensures the stability of the reduced-order model (ROM). We study the potential of the proposed RBM in terms of computational efficiency, accuracy, and storage requirements, and we show that the RBM leads to 100-fold speedups for a two-dimensional case and 1000-fold speedups for a three-dimensional case with an upper frequency of 2 and 1 kHz, respectively. While the FOM simulations needed to construct the ROM are expensive, we demonstrate that the ROM has the potential of being 3 orders of magnitude faster than the FOM when four different boundary conditions are simulated per room surface.

Seamless Integration of the Coastal Ocean in Global Marine Carbon Cycle Modeling

AbstractWe present the first global ocean‐biogeochemistry model that uses a telescoping high resolution for an improved representation of coastal carbon dynamics: ICON‐Coast. Based on the unstructured triangular grid topology of the model, we globally apply a grid refinement in the land‐ocean transition zone to better resolve the complex circulation of shallow shelves and marginal seas as well as ocean‐shelf exchange. Moreover, we incorporate tidal currents including bottom drag effects, and extend the parameterizations of the model's biogeochemistry component to account explicitly for key shelf‐specific carbon transformation processes. These comprise sediment resuspension, temperature‐dependent remineralization in the water column and sediment, riverine matter fluxes from land including terrestrial organic carbon, and variable sinking speed of aggregated particulate matter. The combination of regional grid refinement and enhanced process representation enables for the first time a seamless incorporation of the global coastal ocean in model‐based Earth system research. In particular, ICON‐Coast encompasses all coastal areas around the globe within a single, consistent ocean‐biogeochemistry model, thus naturally accounting for two‐way coupling of ocean‐shelf feedback mechanisms at the global scale. The high quality of the model results as well as the efficiency in computational cost and storage requirements proves this strategy a pioneering approach for global high‐resolution modeling. We conclude that ICON‐Coast represents a new tool to deepen our mechanistic understanding of the role of the land‐ocean transition zone in the global carbon cycle, and to narrow related uncertainties in global future projections.

A vector-to-sequence based multilayer recurrent network surrogate model for history matching of large-scale reservoir

History matching can estimate the parameter of spatially varying geological properties and provide reliable numerical models for reservoir development and management. However, in practice, high-dimension, multiple-solutions and computational cost are key issues that restrict the application of history matching methods. Recently, the combination of deep-learning-based surrogate model and sampling algorithm has been widely studied in history matching to overcome the limitations. Considering that real-world large-scale reservoirs often have hundreds of thousands or even millions of grid-based uncertain parameters, extracting spatial features using convolutional neural networks requires a lot of computational cost and storage requirements. Therefore, in this work, we mainly study how to use the recurrent neural network (RNN) to construct the surrogate model for history matching. Specifically, we propose a multilayer RNN surrogate model based on a vector-to-sequence modeling framework. The multilayer RNN surrogate model with gated recurrent unit (GRU), termed MLGRU, is developed to approximate the mapping from feature vector of geological realizations to the production data. The feature vector is the low-dimensional representation of geological parameter fields after using the re-parameterization method, while production data are the simulation results of historical period. In addition, we design a log-transformation-based windowed normalization (LTWN) method for the production data, which can enhance the learnability and features of production data. The MLGRU model is incorporated into a multimodal estimation of distribution algorithm (MEDA) to formulate a history matching workflow. The hyperparameters and performance of the proposed MLGRU model are analyzed by numerical experiments on a 2D reservoir model. Furthermore, numerical experiments performed on the Brugge benchmark model, a large-scale 3D reservoir model, demonstrated the performance of the proposed surrogate model and history matching method.

Efficient dissipation-preserving scheme for the damped nonlinear Schrödinger equation in three dimensions

We develop an efficient prediction–correction scheme which preserves the exact decay rate of normalization for the damped nonlinear Schrödinger equation in three dimensions. The prediction step only requires solving a scalar nonlinear equation at each grid point and the correction step needs solving one scalar quadratic equation which costs negligible CPU time. Further, compact representation of the prediction–correction scheme saves computational cost and storage requirement greatly. Numerical results are presented to validate the effectiveness and accuracy of the proposed scheme.

A contribution to the conditioning theory of the indefinite least squares problems

The conditioning theory of the indefinite least squares (ILS) problems is considered in the paper. Based on the well acknowledged fact that utilizing the intermediate results produced in the process of solving the problem can largely reduce the computational burden of computing the condition numbers, we employ two hyperbolic matrix decomposition based numerical algorithms to solve the ILS problems, and with its intermediate results the first order perturbation analysis and condition number theory of the ILS problems are fully investigated. The first order perturbation analysis gives a precise characterization of the matrix factors in influencing the forward errors. For the new explicit expressions of the condition numbers consisting of intermediate results, some compact forms or upper bounds are proposed to reduce its computational burden and storage requirement. Moreover, we also devise some efficient numerical algorithms to estimate the condition numbers by exploiting its special structures. Numerical experiments are given to illustrate the efficiency of our results.