Multiple Parallel Processing Research Articles

Multi-scale time series analysis using TT-ConvLSTM technique for bearing remaining useful life prediction

This study advocates the utilization of a parallel neural network (PNN) architecture for the estimation of remaining useful life (RUL) of rolling element bearings. Conventional machine learning methods have often fallen short in processing machinery sensor data efficiently and accurately, given its vast nature and the temporal dynamics of component health conditions. To address this limitation, a robust PNN architecture is proposed which incorporates multiple parallel processing pathways. These pathways are equipped with diverse input neurons, capturing data from mechanical component condition sensors. The output neuron then efficiently predicts the RUL, capturing the degradation stages throughout the component's lifespan. The PNN structure provides better accuracy and computation time by efficiently handling huge amounts of data simultaneously and integrating both spatial and temporal information present. Additionally, time-transformer and recurrent neural networks (RNNs) are used to handle complex time series data. Improvement methodologies, such as positional encoding combined with a multi-headed self-attention mechanism and the convolutional long short-term memory (ConvLSTM) neural network, are employed. These methodologies adeptly handle spatiotemporal dependencies inherent in multidimensional features extracted across the time domain, frequency domain, and time–frequency domain, thereby considerably boosting the model's efficacy. Two case studies are conducted on XJTU-SY rolling element bearing dataset and FEMTO bearing dataset to validate and display the generalization capabilities of the proposed methodology, where PNN performed exceptionally in terms of accuracy and efficiency. It is concluded that with a novel implementation of variational stride temporal window strategy and utilizing the full capabilities of a PNN, the RUL of bearings can be predicted very accurately. The algorithm strategy is very robust and can be applied to other machine components as well. Subsequently, the code for the algorithm will be open-sourced.

Parallel river channel routing computation based on a straightforward domain decomposition of river networks

River routing is an important component of hydrological models, and is a time-consuming task since it is usually based on finer-scale river channel networks and runs with shorter time steps. In order to utilize the parallel computing capabilities of today’s computers, it is necessary to decompose river networks into multiple domains and allocate them to multiple parallel processor cores to process. In this study, a straight-forward domain decomposition method is proposed for an efficient Muskingum routing algorithm on an arbitrary river network of tree-like hierarchical structure: evenly split the upstream-to-downstream sorted river segments into multiple blocks (domains). Within each domain, the cross-domain segments (CDSs), which are defined as those segments receiving inflows from other domains, need to be identified. For such domain decomposition, three parallelization schemes (the standard scheme, the full-parallelization scheme and the simplified scheme) of the Muskingum river routing are proposed. The total number of CDSs increases with the increasing number of domains. For the standard scheme, when too many CDSs are produced, the computation time that saved by the parallelization in its first step should be canceled out by the serial processing in its second step. The sorting order of river segments also affects execution speed and output errors: a depth-first traverse generates much fewer CDSs than a width-first traverse. The number of CDSs were analyzed on decomposing the networks of the Yellow River and the Mississippi, and the optimal number of domains is close to 16 (Yellow River, 22′750 river segments) and 32(Mississippi, 64′579 river segments), respectively. The computing time over different numbers of domains was also tested, based on the parallel execution by both multicore CPU and GPU. The errors produced by the full-parallelization scheme and the simplied scheme are relatively small, especially for the output at the mainstem segments. The parallel river routing can correctly produce the result as expected, and the application to actual river flow modeling produces reasonable outputs.

All-in-One Preparation Strategy Integrated in a Miniaturized Device for Fast Analyses of Biomarkers in Biofluids by Surface Enhanced Raman Scattering.

Complex and tedious sample preparation processes have greatly limited rapid analyses of biological samples. In this work, an all-in-one sample preparation strategy based on a miniaturized gas membrane separation/oven ring enrichment (GMS/ORE) device was developed for efficient surface enhanced Raman scattering (SERS) analyses of trace biomarkers in biofluid samples. This strategy integrating gasification separation, liquid trapping, derivatization SERS activation, and coffee-ring enrichment could highly promote the efficiency of sample preparation. Meanwhile, the edges of membranes modified by the hydrophobic-infusing slippery liquid-induced uniform "coffee-ring" effect could significantly improve the sensitivity and stability for SERS quantification. By adapting proper derivatization approaches to the miniaturized GMS/ORE pretreatment, the matrix effects in samples could be prominently eliminated, and clear SERS responses could be obtained for the selective analyses of target biomarkers. The miniaturized GMS/ORE device was practically applied for SERS analyses of trace biomarkers in biofluids, including hydrogen sulfide in saliva samples, creatinine in serum samples, and sarcosine, creatinine, and dimethyl disulfide in urine samples. Accurate quantification of all biomarkers was achieved with recoveries of 89.5%-120.0%, and the contents found by GMS/ORE-SERS matched well with those found by corresponding chromatographic methods with relative errors from -8.6% to 9.3%. The miniaturized GMS/ORE device with multiple parallel processing units could simultaneously treat eight samples in one run with a total analysis time of 40 min. Such an efficient all-in-one strategy integrated on a miniaturized device possesses great potential for fast on-site/point-of-care detection in analytical science and clinical medicine.

Modified Gamma Network: Design and Reliability Evaluation

Background: VLSI technology advancements have resulted the requirements of high computational power, which can be achieved by implementing multiple processors in parallel. These multiple processors have to communicate with their memory modules by using Interconnection Networks (IN). Multistage Interconnection Networks (MIN) are used as IN, as they provide efficient computing with low cost. Objective: The objective of the study is to introduce new reliable Gamma MIN named as a Modified Gamma Interconnection Network (MGIN), which provide reliability and fault-tolerance with less number of stages of Switching element only. Methods: Switching Element (SE) of bigger size i.e. 2×3/3×2 has been employed at input/output stages inspite of 1×3/3×1 sized SE at input/output stages with reduction in one intermidiate stage. Fault tolerance has been introduced in the form of disjoint paths formed between each sourcedestnation node pair. Hence reliability has been improved. Results: Terminal, Broadcast and Network Reliability has been evaluated by using Reliability Block Diagrams for each source-destination node pair. The results have been shown, which depicts the higher reliability values for newly proposed network. The cost analysis shows that new MGIN is a cheaper network than other Gamma variants. Conclusion: MGIN has better reliability and Fault-tolerance than priviously proposed Gamma MIN.

A multigrid local smoother approach for a domain decomposition solver over non‐matching grids

AbstractIn this paper we consider a multigrid approach for solving elliptic equations over non‐matching grids with domain decomposition methods. The domain is partitioned into subdomains with different mesh levels that do not match at the interface. The proposed algorithm searches for the global solution over different levels by projecting the residuals on the overlap region. This method is used in conjunction with a domain decomposition solver which only requires, in each iteration step, the solutions of several small local subproblems over finite element blocks. This algorithm is shown to converge to the solution of the corresponding Lagrange multiplier problem for non‐matching grids. The convergence properties of the algorithms are analyzed and numerical examples are presented. When the multigrid and domain decomposition approaches are combined, the method is shown to be reliable and easy to implement. Furthermore the local nature of the solver allows for a straightforward implementation on multiple parallel computers and graphics processing unit (GPU) clusters.

An efficient parallel iteration algorithm for nonlinear diffusion equations with time extrapolation techniques and the Jacobi explicit scheme

To numerically simulate the radiation diffusion problem with high parallel efficiency, we present a new parallel iteration algorithm for nonlinear diffusion equations. The algorithm is based on the domain decomposition method, and it integrates time extrapolation techniques and an advanced Jacobi explicit scheme. The domain decomposition method decomposes a large global problem into multiple sub-problems which can be solved on multiple processors in parallel. The time extrapolation technique gives the prediction values with the second or third order accuracy in time for the current time layer by specific combinations of the previous two or three time layers. The advanced Jacobi explicit scheme further improves the precision of the prediction values. Overall, the proposed algorithm makes the prediction values of the inner boundary more reasonable and offers accurate iterative initial values for all cells, which reduces the number of nonlinear iterations and improves the parallel computation efficiency remarkably.

A new bifuzzy optimization method for remanufacturing scheduling using extended discrete particle swarm optimization algorithm

As the energy shortage and environment pollution become increasingly deteriorating, remanufacturing has become a hot research field for its energy-saving and environmental advantages. In remanufacturing systems, scheduling is a key step that directly affects the successful realization of remanufacturing benefits. Since the remanufacturing involves many two-fold uncertainties, scheduling for remanufacturing is more challenging than for traditional manufacturing. However, few studies have addressed these two-fold uncertainties in remanufacturing environments. Thus, this study proposes a new bifuzzy remanufacturing scheduling model, which handles many two-fold uncertainties in remanufacturing and applies the bifuzzy variable to describe these uncertainties. Multiple non-identical parallel processing lines for remanufacturing end-of-life products have also been integrated into the proposed model. An extended discrete particle swarm optimization algorithm with an effective representation scheme is proposed to solve the model. In the presented algorithm, three new guiding directions and a new multi-directional guiding strategy are designed to enhance the diversity of population, and a new position updating mechanism, local search strategy, and anti-stagnation mechanism are integrated to improve the algorithmic performance. Experiments are performed to verify the effectiveness of the proposed model by comparing it with the traditional deterministic model, and the effectiveness of the proposed algorithm in solving this model by comparing it with other baseline algorithms.

Parallel Multiphysics Simulation of Package Systems Using an Efficient Domain Decomposition Method

With the continuing downscaling in feature sizes, the thermal impact on material properties and geometrical deformations can no longer be ignored in the analysis of the electromagnetic compatibility or electromagnetic interference of package systems, including System-in-Package and antenna arrays. We present a high-performance numerical simulation program that is intended to perform large-scale multiphysics simulations using the finite element method. An efficient domain decomposition method was developed to accelerate the multiphysics loops of electromagnetic–thermal stress simulations by considering the fact that the electromagnetic field perturbations caused by geometrical deformation are small and constrained in one or a few subdomains. The multi-level parallelism of the algorithm was also obtained based on an in-house developed parallel infrastructure. Numerical examples showed that our algorithm is able to enable simulation with multiple processors in parallel and, more importantly, achieve a significant reduction in computation time compared with traditional methods.

Short term power dispatch using neural network based ensemble classifier

This work reports the application of ensemble artificial neural networks, a machine learning technique, in the solution of short term optimal power generation on the IEEE 30-bus test system. The study, carried out using MATLAB, is being reported. The motive for using ensemble artificial neural networks has been to take advantage of multiple parallel processors computing rather than the traditional serial computation. The Bootstraps are obtained through small bags in the Bagging algorithm and are combined by averaging. Employing ensemble neural networks reduces bias and variance in machine learning so that the overall error is reduced, and hence better prediction can be achieved. The results obtained from the proposed algorithm are compared with those from the other heuristic and conventional techniques to validate the effectiveness of the proposed methodology.

SEGIN-Minus: A New Approach to Design Reliable and Fault-Tolerant MIN

Background:VLSI technology advancements have resulted the requirements of high computational power, which can be achieved by implementing multiple processors in parallel. These multiple processors have to communicate with their memory modules by using Interconnection Networks (IN). Multistage Interconnection Networks (MIN) are used as IN, as they provide efficient computing with low cost.Objective:the objective of the study is to introduce new reliable MIN named as a (Shuffle Exchange Gamma Interconnection Network Minus) SEGIN-Minus, which provide reliability and faulttolerance with less number of stages.Methods:MUX at input terminal and DEMUX at output terminal of SEGIN has been employed with reduction in one intermidiate stage. Fault tolerance has been introduced in the form of disjoint paths formed between each source-destnation node pair. Hence reliability has been improved.Results:Terminal, Broadcast and Network Reliability has been evaluated by using Reliability Block Diagrams for each source-destination node pair. The results have been shown, which depicts the hiher reliability values for newly proposed network. The cost analysis shows that new SEGINMinus is a cheaper network than SEGIN.Conclusion:SEGIN-Minus has better reliability and Fault-tolerance than priviously proposed SEGIN.

Low-Complexity PAPR Reduction Method for OFDM Systems Based on Real-Valued Neural Networks

High peak-to-average power ratio (PAPR) in orthogonal frequency division multiplexing (OFDM) systems is one of the major drawbacks in wireless transmitters. In this letter, a novel low complexity PAPR reduction method based on the real-valued neural network (NN) is proposed. This method first builds the PAPR reduction module using the real-valued NN to achieve the reduction of PAPR. To reconstruct the transmission signal, the PAPR decompression module is introduced into the receiving end to recover the signal in order to minimize the bit error rate (BER) of the systems. The PAPR reduction module and the PAPR decompression module can be jointly trained offline, therefore, the reduction of PAPR and the minimization of BER can be achieved simultaneously. To reduce the process rate of the PAPR reduction, the trained model uses multiple parallel processing links to achieve the reduction of the PAPR. The extensive simulations indicate that the proposed method outperforms previously available methods in terms of PAPR and BER, while having low complexity.

Density based clustering and fuzzy clustering for efficient clustering of big data in hadoop ecosystem

In this paper suggesting to use parallel distributed Hadoop Map Reduce technology. Map Reduce is a parallel technology which can process bigdata by creating multiple instances of thread. Map will take input data and split it into multiple chunks or parts and distribute all parts to different reducers. All reducers will process the data and send result back to mapper. Mapper gather output from all mappers and then generate a single output. Due to multiple parallel processing of Map Reduce technology allow us to process any amount of data.In this paper author describing clustering algorithms such Density Based Clustering and Fuzzy Clustering. Both algorithms are not efficient to group all similar data to single cluster and make some data to compromise by putting little un-similar to different clusters. To implement this project author is using National Climatic Data Center (NCDC) dataset which contains climate information. To find out similar temperature on different dates author is applying Hybrid clustering algorithm. From this dataset author is using date and temperature value and then passing date as key to Mapper and temperature as value to Mapper. Mapper always read data in the form of key value pairs. In dataset date we can find at position 6 to14 and temperature we can find at position 39 to 45. Below are some dataset example values.

DSS approach for heterogeneous parallel machines scheduling considering proximate supply chain constraints

This paper describes the basis of a Decision Support System (DSS) designed to schedule fertiliser production orders to be delivered within time windows, in plants made up of multiple heterogeneous parallel processors (production lines), considering that fertiliser production rates and nomenclatures depend on lines, that setup times depend on sequence and lines, and taking into account downtime constraints (preventive maintenance …). A mixed linear programming model is encapsulated in the DSS which considers the schedule’s impacts, immediately upstream and downstream of plants in the supply chain. These side-effects may make the proposed solution unfeasible and the DSS helps redefining the problem to avoid them.

An Optimization Method for Hologram Generation on Multiple GPU-based Parallel Processing

An Optimization Method for Hologram Generation on Multiple GPU-based Parallel Processing

Optimal Power Generation in Energy-Deficient Scenarios Using Bagging Ensembles

This paper presents an improved technique for optimal power generation using ensemble artificial neural networks (EANN). The motive for using EANN is to benefit from multiple parallel processor computing rather than traditional serial computation to reduce bias and variance in machine learning. The load data is obtained from the load regulation authority of Pakistan for 24 hours. The data is analyzed on an IEEE 30-bus test system by implementing two approaches; the conventional artificial neural network (ANN) with feed-forward back-propagation model and a Bagging algorithm. To improve the training of ANN and authenticate its result, first the Load Flow Analysis (LFA) on IEEE 30 bus is performed using Newton Raphson Method and then the program is developed in MATLAB using Lagrange relaxation (LR) framework to solve a power-generator scheduling problem. The bootstraps for the EANN are obtained through a disjoint partition Bagging algorithm to handle the fluctuating power demand and is used to forecast the power generation. The results of MATLAB simulations are analyzed and compared along with computational complexity, therein showing the dominance of the EANN over the traditional ANN strategy that closed to LR.

ARC: adaptive ray-tracing with CUDA, a new ray tracing code for parallel GPUs

We present the methodology of a photon-conserving, spatially-adaptive, ray-tracing radiative transfer algorithm, designed to run on multiple parallel Graphic Processing Units (GPUs). Each GPU has thousands computing cores, making them ideally suited to the task of tracing independent rays. This ray-tracing implementation has speed competitive with approximate momentum methods, even with thousands of ionization sources, without sacrificing accuracy and resolution. Here, we validate our implementation with the selection of tests presented in the "cosmological radiative transfer codes comparison project," to demonstrate the correct behavior of the code. We also present a selection of benchmarks to demonstrate the performance and computational scaling of the code. As expected, our method scales linearly with the number of sources and with the square of the dimension of the 3D computational grid. Our current implementation is scalable to an arbitrary number of nodes possessing GPUs, but is limited to a uniform resolution 3D grid. Cosmological simulations of reionization with tens of thousands of radiation sources and intergalactic volumes sampled with 1024$^3$ grid points take about 30 days on 64 GPUs to reach complete reionization.

Application of distributed auction to multi-UAV task assignment in agriculture

Abstract: This paper addresses the multiple farming task assignment for agricultural UAVs by presenting a decentralized auction algorithm, where computation and information are distributed among multiple parallel processing units (UAVs). The scheme iterates between bundle construction phase and two conflict resolution phases, and then converges to a task allocation and route plan simultaneously. In the bundle construction stage, each UAV groups the tasks with commonalities by considering its own capacities of flight endurance, weight load, battery, data storage, etc. In the following conflict resolution stages, the winning UAVs for tasks are determined by the information exchanged between UAVs. Later, the proposed algorithm is shown to have a low demand for communication and is proved to be able to achieve 50% optimality. Finally, an application of collecting the data of ground sensor nodes (including moving nodes) is used to assess the performance of the proposed scheme. Numerical experiments confirm superior convergence properties and performance of the proposed algorithm when compared with existing task-allocation algorithms. Keywords: unmanned aerial vehicle, farming task allocation, route plan, distributed, auction,agricultural aviation DOI: 10.33440/j.ijpaa.20180101.0008 Citation: Hu J, Yang J C. Application of distributed auction to multi-UAV task assignment in agriculture. Int J Precis Agric Aviat, 2018; 1(1): 44–50.

Automatic Fitting of Binary Interaction Parameters for Multi-fluid Helmholtz-Energy-Explicit Mixture Models

In the highest-accuracy mixture models available today, these being the multi-fluid Helmholtz-energy-explicit formulations, there are a number of binary interaction parameters that must be obtained through correlation or estimation schemes. These binary interaction parameters are used to shape the thermodynamic surface and yield higher-fidelity predictions of various thermodynamic properties including vapor-liquid equilibria and homogeneous p-v-T data, among others. In this work, we have used a novel and entirely automatic evolutionary optimization algorithm written in the python programming language to fit the two most important interaction parameters for more than 1100 binary mixtures. This fitting algorithm can be run on multiple processors in parallel, resulting in a reasonable total running time for this large set of binary mixtures. For more than 830 of the binary pairs, the median absolute relative error in bubble-point pressure is less than 5%. The source code for the fitter is provided as supplem...

Approximate partitioned method of snapshots for POD

The proper orthogonal decomposition (POD) method has been widely used to construct efficient numerical surrogate models for computationally intensive applications in control and optimization. An inherent challenge with this method is that POD basis generation can be computationally expensive due to the huge size of the input snapshot data obtained from typical high-fidelity, large-scale dynamic system simulations. However, if the process can be distributed into much smaller tasks over multiple processors in parallel, computational time can be drastically reduced. In this paper, we put forth a novel partitioned method for generating the POD basis from snapshot data. This method preserves the distributed nature of the data and takes advantage of parallelism for computation. Additionally, it greatly reduces subtask communication volume. Two numerical examples are presented that demonstrate the effectiveness of the new method.

Fully Sequential Procedures for Large-Scale Ranking-and-Selection Problems in Parallel Computing Environments

Fully sequential ranking-and-selection (R&S) procedures to find the best from a finite set of simulated alternatives are often designed to be implemented on a single processor. However, parallel computing environments, such as multi-core personal computers and many-core servers, are becoming ubiquitous and easily accessible for ordinary users. In this paper, we propose two types of fully sequential procedures that can be used in parallel computing environments. We call them vector-filling procedures and asymptotic parallel selection procedures, respectively. Extensive numerical experiments show that the proposed procedures can take advantage of multiple parallel processors and solve large-scale R&S problems.