Master-slave Parallel Research Articles

A ten time smaller version of CPC global daily precipitation dataset for parallel distributed processing in Matlab and R

A ten times smaller version of CPC Global Unified Gauge-Based Daily Precipitation Dataset (1979–2021) is released and described in this paper. Its usability is tested and proved, firstly, by illustrating that the transition to the derived smaller dataset is a case of Pearson correlation transitivity starting with the scale of global yearly data and, secondly, by using the original correlation performance criterion that the original dataset satisfies relative to the set of global actual measurements on the record. Subsequently, the daily, weekly, and monthly data cases are considered and discussed. The dataset is (re)structured for parallel processing in Matlab and R from the level of global daily data. Considering the above arguments and its reduced size, the derived dataset is appropriate to be used with Matlab and R as a replacement for the original dataset, especially for the case when much faster exploratory master-slave parallel and distributed Matlab and/or R tasks will run over locally distributed data on the slaves.

RETRACTED ARTICLE: Effectiveness of Mixed Fuzzy Time Window Multi-objective Allocation in E-Commerce Logistics Distribution Path

The study of logistics distribution network under e-commerce environment is conducive to the establishment of efficient logistics distribution system, but also to promote the further development of e-commerce and improve social benefits of great significance. This study considers multiple fuzzy factors and introduces a customer fuzzy time window with variable coefficients, establishes a multi-objective set allocation integrated multi-level location path planning model, and proposes an archive type multi-objective simulated annealing improvement algorithm based on master–slave parallel framework embedded taboo search to solve the model. Tabu search and large-scale neighborhood algorithm are used to solve the initial solutions of the first level network and the second level network respectively, and archival reception criterion is introduced to deal with the multi-objective problem. The results of the proposed algorithm for the two-level site-routing problem are less than 6% different from the internationally known optimal solution. The master–slave parallel computing framework improves the efficiency of the algorithm by about 6.38%. The experimental results prove the effectiveness and necessity of the improved optimization. In addition, this study simulates the site-routing problem model constructed by the study by extending the data of standard examples. The experimental results prove the correctness and reference significance of the multilevel site-routing problem model with multiple fuzzy factors.

Distributed Tree-Based Machine Learning for Short-Term Load Forecasting With Apache Spark

Machine learning algorithms have been intensively applied to perform load forecasting to obtain better accuracies as compared to traditional statistical methods. However, with the huge increase in data size, sophisticated models have to be created which require big data platforms. Optimal and effective use of the available computational resources can be attained by maximizing the effective utilization of the cluster nodes. Parallel computing is demanded to allow for optimal resource utilization in dealing with smart grid big data. In this paper, a master-slave parallel computing paradigm is utilized and experimented with for load forecasting in a multi-AMI environment. The paper proposes a concurrent job scheduling algorithm in a multi-energy data source environment using Apache Spark. An efficient resource utilization strategy is proposed for submitting multiple Spark jobs to reduce job completion time. The optimal value of clustering is used in this paper to cluster the data into groups to be able to reduce the computational time additionally. Multiple tree-based machine learning algorithms are tested with parallel computation to evaluate the performance with tunable parameters on a real-world dataset. One thousand distribution transformers’ real data from Spain for three years are used to demonstrate the performance of the proposed methodology with a trade-off between accuracy and processing time.

Optimum State-of-Charge Operating Range for Frequency Regulation of Energy Storage Systems Using a Master–Slave Parallel Genetic Algorithm

Lithium batteries are used for frequency regulation in power systems because of their fast response and high efficiency. Lithium batteries have different life characteristics depending on their type, and it is necessary to set the optimal state-of-charge (SOC) operating range considering these characteristics to obtain the maximum gain. In general, narrowing the operating range increases the service life but may lower the performance of charging and discharging operations in response to frequency fluctuations, and vice versa. We present performance assessment indicators that consider charging and discharging due to frequency variations and lifespan of the batteries. However, to evaluate the performance, while reflecting the non-linear life characteristics of lithium batteries, simulating the entire operation is necessary, which requires a long calculation time. Therefore, we propose a master–slave parallel genetic algorithm to derive the optimal SOC operating range with reduced calculation time. A simulation program was implemented to evaluate the computational performance that determines the optimal SOC range. The proposed method reduces the calculation time while considering the non-linear life characteristics of lithium batteries. It was confirmed that a more accurate SOC operating range could be calculated by simulating the entire life span.

Optimal Walker Constellation Design of LEO-Based Global Navigation and Augmentation System

Low Earth orbit (LEO) satellites located at altitudes of 500 km~1500 km can carry much stronger signals and move faster than medium Earth orbit (MEO) satellites at about a 20,000 km altitude. Taking advantage of these features, LEO satellites promise to make contributions to navigation and positioning where global navigation satellite system (GNSS) signals are blocked as well as the rapid convergence of precise point positioning (PPP). In this paper, LEO-based optimal global navigation and augmentation constellations are designed by a non-dominated sorting genetic algorithm III (NSGA-III) and genetic algorithm (GA), respectively. Additionally, a LEO augmentation constellation with GNSS satellites included is designed using the NSGA-III. For global navigation constellations, the results demonstrate that the optimal constellations with a near-polar Walker configuration need 264, 240, 210, 210, 200, 190 and 180 satellites with altitudes of 900, 1000, 1100, 1200, 1300, 1400 and 1500 km, respectively. For global augmentation constellations at an altitude of 900 km, for instance, 72, 91, and 108 satellites are required in order to achieve a global average of four, five and six visible satellites for an elevation angle above 7 degrees with one Walker constellation. To achieve a more even coverage, a hybrid constellation with two Walker constellations is also presented. On this basis, the GDOPs (geometric dilution of precision) of the GNSS with and without an LEO constellation are compared. In addition, we prove that the computation efficiency of the constellation design can be considerably improved by using master–slave parallel computing.

A Parallel Algorithm Framework for Feature Extraction of EEG Signals on MPI

In this paper, we present a parallel framework based on MPI for a large dataset to extract power spectrum features of EEG signals so as to improve the speed of brain signal processing. At present, the Welch method has been wildly used to estimate the power spectrum. However, the traditional Welch method takes a lot of time especially for the large dataset. In view of this, we added the MPI into the traditional Welch method and developed it into a reusable master-slave parallel framework. As long as the EEG data of any format are converted into the text file of a specified format, the power spectrum features can be extracted quickly by this parallel framework. In the proposed parallel framework, the EEG signals recorded by a channel are divided into N overlapping data segments. Then, the PSD of N segments are computed by some nodes in parallel. The results are collected and summarized by the master node. The final PSD results of each channel are saved in the text file, which can be read and analyzed by Microsoft Excel. This framework can be implemented not only on the clusters but also on the desktop computer. In the experiment, we deploy this framework on a desktop computer with a 4-core Intel CPU. It took only a few minutes to extract the power spectrum features from the 2.85 GB EEG dataset, seven times faster than using Python. This framework makes it easy for users, who do not have any parallel programming experience in constructing the parallel algorithms to extract the EEG power spectrum.

Parallel Regional Segmentation Method of High-Resolution Remote Sensing Image Based on Minimum Spanning Tree

With finer spatial scale, high-resolution images provide complex, spatial, and massive information on the earth’s surface, which brings new challenges to remote sensing segmentation methods. In view of these challenges, finding a more effective segmentation model and parallel processing method is crucial to improve the segmentation accuracy and process efficiency of large-scale high-resolution images. To this end, this study proposed a minimum spanning tree (MST) model integrated into a regional-based parallel segmentation method. First, an image was decomposed into several blocks by regular tessellation. The corresponding homogeneous regions were obtained using the minimum heterogeneity rule (MHR) partitioning technique in a multicore parallel processing mode, and the initial segmentation results were obtained by the parallel block merging method. On this basis, a regionalized fuzzy c-means (FCM) method based on master-slave parallel mode was proposed to achieve fast and optimal segmentation. The proposed segmentation approach was tested on high-resolution images. The results from the qualitative assessment, quantitative evaluation, and parallel analysis verified the feasibility and validity of the proposed method.

Traffic Network Flow Prediction Using Parallel Training for Deep Convolutional Neural Networks on Spark Cloud

Traffic flow in a road network is mutually interactive and interdependent with each other. It is challenging to describe the dynamics of traffic network flow by using analytical methods. In this article, the deep convolutional neural network (DCNN) model is employed to address traffic network flow prediction. To improve the parameter learning efficiency confronting traffic big data, a parallel training approach is developed for the DCNN prediction model. The theoretical foundation is developed for the parallel training algorithm of the DCNN model. A master-slave parallel computing solution for traffic network flow prediction is implemented on the Spark cloud. Real data of traffic network flow are applied to verify the effectiveness of the DCNN prediction model and the parallel training algorithm. The experimental results demonstrate that the DCNN prediction model for traffic network flow outperforms the typical prediction models based on backpropagation neural networks, support vector regressions, radial basis functions, and decision tree regressions. The proposed parallel training method can improve the training efficiency and obtain global features of the entire dataset from local learning with regard to the respective data subsets.

Parallel modeling of cell suspension flow in complex micro-networks with inflow/outflow boundary conditions

We proposed a parallel framework with inflow/outflow boundary conditions for particle-based simulations of cell suspension flow through complex micro-networks that is often encountered in microcirculation and microfluidic devices. The behaviors of fluid and cells are modeled by a hybrid approach of smoothed dissipative particle dynamics (SDPD) and immersed boundary method (IBM). The simulation system is associated with two more domains, the inflow domain at the inlets and the outflow domains at the outlets, to implement the inflow/outflow boundary conditions. The inflow is duplicated into the system from a fully-developed flow generated in the inflow domain under a periodic boundary condition. The outflow is iteratively controlled by two adaptive forces to maintain mass and momentum conservation, and meanwhile, the fluid and cells leaving the outlets are removed from the system. A master-slave parallel configuration is constructed, where the master thread is responsible for dividing the simulation domain or allocating the simulation tasks to three types of slave threads, cell, inflow and hybrid slave threads. To save message communication, each cell is assigned to a separate thread by task decomposition, while the inflow/hybrid slave threads are divided by domain decomposition. Four validation studies were carried out for the fluid flows in a straight and a bifurcated tubes, the deformation of a capsule in a straight tube, as well as the rheology of many red blood cells (RBCs) in a straight tube. The results were compared with their corresponding analytical solutions or previously-published numerical results, and good agreements were observed on the velocity field, cell deformation, and rheological behaviors of cells. Moreover, two numerical experiments were conducted to demonstrate the capability of the proposed methodology. One is RBCs through a microvascular network with an inlet but four outlets, and the other is those through a serpentine microfluidic chip with several complex geometric features, including bifurcations, sharp corners, sudden expansion and multiple outlets. Both of these experiments not only demonstrated the capability of the new methodology, but also suggested a potential to simulate real-world biomedical problems, such as thrombus formation, tumor metastasis, drug delivery, and so on.

A Master Slave Parallel Genetic Algorithm for Feature Selection in High Dimensional Datasets

Feature Selection in High Dimensional Datasets is a combinatorial problem as it selects the optimal subsets from N dimensional data having 2N possible subsets. Genetic Algorithms are generally a good choice for feature selection in large datasets, though for some high dimensional problems it may take varied amount of time - few seconds, few hours or even few days. Therefore, it is important to use Genetic Algorithms that can give quality results in reasonably acceptable time limit. For this purpose, it is becoming necessary to implement Genetic Algorithms in an efficient manner. In this paper, a Master Slave Parallel Genetic Algorithm is implemented as a Feature Selection procedure to diminish the time intricacies of sequential genetic algorithm. This paper describes the speed gains in parallel Master-Slave Genetic Algorithm and also discusses the theoretical analysis of optimal number of slaves required for an efficient master slave implementation. The experiments are performed on three high-dimensional gene expression data. As Genetic Algorithm is a wrapper technique and takes more time to find the importance of any feature, Information Gain technique is used first as pre-processing task to remove the irrelevant features.

Analysis of semi-asynchronous multi-objective evolutionary algorithm with different asynchronies

This paper proposes a novel master–slave parallel evolutionary algorithm (EA) approach with different asynchrony and provides its detailed analyses on multi-objective optimization problems. We express the proposed EA with different asynchrony as a semi-asynchronous EA. A semi-asynchronous EA generates new solutions whenever evaluations of the predefined number of solutions complete, unlike a conventional synchronous EA waits for evaluations of all solutions to generate the next population. To establish a semi-asynchronous EA, this paper introduces an asynchrony parameter that is used to decide how many solutions are waited to generate new solutions. We conduct an experiment to verify the effectiveness of the proposed semi-asynchronous EA on benchmark problems with the several variances of the evaluation time. In the experiment, we apply a semi-asynchronous EA to NSGA-II and NSGA-III, which are well-known multi-objective EAs. The semi-asynchronous NSGA-IIs and the semi-asynchronous NSGA-IIIs with different asynchronies are compared on multi-objective optimization benchmark problems. The experimental result reveals that semi-asynchronous approaches with an appropriate asynchrony have possibility to outperform the asynchronous and the synchronous ones. Additionally, detailed analysis reveals that an appropriate asynchrony may vary not only depends on a target problem but also depends on the degree of the evolution process.

Revisiting the Parallel Strategy for DOACROSS Loops

DOACROSS loops are significant parts in many important scientific and engineering applications, which are generally exploited pipeline/wave-front parallelism by loop transformations. However, previous work almost statically performs iterations in parallel threads, thus causing a waste of computing resources in thread synchronization. This paper proposes a brand-new parallel strategy for DOACROSS loops that provides a dynamic task assignment with reduced dependences to achieve wave-front parallelism through loop tiling. The proposed strategy uses a master-slave parallel mode and some customized structures to realize dynamic and flexible parallelization, which effectively avoids threads from waiting in communication. An efficient tile size selection (TSS) approach is also proposed to preserve data reuse in cache for tiled codes. The experimental results show that the proposed parallel strategy obtains good and stable speedups over six typical benchmarks with different problem sizes and different numbers of threads on an Intel® Xeon® 32-core server. And it outperforms two static strategies, a barrier-based strategy and a post/wait-based strategy, by 32% and 20% in average performance, respectively. This strategy also yields a better performance than a mutex-based dynamic strategy. Besides, it has been demonstrated that the proposed TSS approach can achieve a near-optimal performance and is comparable with a state-of-the-art TSS approach.

Parallel computing method of deep belief networks and its application to traffic flow prediction

Deep belief networks (DBNs) with outstanding advantages of learning input data features have attained particular attention and are applied widely in image processing, speech recognition, natural language interpretation, disease diagnosis, among others. However, owing to large data, the training processes of DBNs are time-consuming and may not satisfy the requirements of real-time application systems. In this study, a single dataset is decomposed into multiple subdatasets that are distributed to multiple computing nodes. Multiple computing nodes learn the features of their own subdatasets. On the precondition of the remaining features where one computing node learns from the total dataset, the single dataset learning models and algorithms are extended to the cases where multiple computing nodes learn multiple subdatasets in a parallel manner. Learning models and algorithms are proposed for the parallel computing of DBN learning processes. A master–slave parallel computing structure is designed, where the slave computing nodes learn the features of their respective subdatasets and transmit them to the master computing node. The master computing node is critical in synthesizing the learned features from the respective slave computing nodes. The broadcast, synchronization, and synthesis are repeated until all features of subdatasets have been learned. The proposed parallel computing method is applied to traffic flow prediction using practical traffic flow data. Our experimental results verify the effectiveness of the parallel computing method of DBN learning processes in terms of decreasing pre-training and fine-tuning times and maintaining the prominent feature learning abilities.

A probabilistic multi-objective approach for power flow optimization in hybrid wind-PV-PEV systems

This paper formulates and solves a probabilistic optimal power flow approach (POPF) for a hybrid power system that includes plug-in electric vehicles (PEV), photovoltaic (PV) and wind energy (WE) sources. In the proposed approach, the Monte Carlo Simulation (MCS) was combined with the antithetic variates method (AVM) to determine the probability distribution function (PDF) of the power generated by the hybrid system. To reduce the computational cost of the optimal power flow calculations, we solved the POPF problem using a master–slave parallel epsilon variable multi objective genetic algorithm (Pev-MOGA). The performance of the proposed approach was assessed using the IEEE 30-bus, 57-bus and 118-bus power systems. Various scenarios incorporating several configurations of WEs, PVs and PEVs sources were considered in the evaluation. Sensitivity analysis was also performed for further assessment. The obtained results along with a comparison analysis with other optimization algorithms confirmed the effectiveness of the proposed approach in accurately providing a set of optimal solutions for the hybrid power system.

A lightweight plug-and-play elasticity service for self-organizing resource provisioning on parallel applications

Today cloud elasticity can bring benefits to parallel applications, besides the traditional targets including Web and critical-business demands. This consists in adapting the number of resources and processes at runtime, so users do not need to worry about the best choice for them beforehand. To accomplish this, the most common approaches use threshold-based reactive elasticity or time-consuming proactive elasticity. However, both present at least one problem related to the need of a previous user experience, lack on handling load peaks, completion of parameters or design for a specific infrastructure and workload setting. In this context, we developed a hybrid elasticity service for master–slave parallel applications named Helpar. The proposal presents a closed control loop elasticity architecture that adapts at runtime the values of lower and upper thresholds. The main scientific contribution is the proposition of the Live Thresholding (LT) technique for controlling elasticity. LT is based on the TCP congestion algorithm and automatically manages the value of the elasticity bounds to enhance better reactiveness on resource provisioning. The idea is to provide a lightweight plug-and-play service at the PaaS (Platform-as-a-Service) level of a cloud, in which users are completely unaware of the elasticity feature, only needing to compile their applications with Helpar prototype. For evaluation, we used a numerical integration application and OpenNebula to compare the Helpar execution against two scenarios: a set of static thresholds and a non-elastic application. The results present the lightweight feature of Helpar, besides highlighting its performance competitiveness in terms of application time (performance) and cost (performance × energy) metrics.

Efficient parallel evolutionary algorithms for deadline-constrained scheduling in project management

Deadline-constrained scheduling in project management is a NP-hard optimisation problem with major relevance in software engineering and other real-life situations dealing with the planning of activities that must be completed before specific dates. This article introduces efficient parallel versions for two evolutionary algorithms genetic algorithm and hybrid evolutionary algorithm, to solve the deadline-constrained scheduling problem in project management. The proposed algorithms have been engineered to compute accurate solutions in reduced execution times. Specific evolutionary operators, including a parallel local search operator in the hybrid evolutionary algorithm, are proposed for efficiently solving realistic problem instances, and both a master-slave parallel strategy and a distributed subpopulation model are applied to further improve the computational efficiency and the results quality. The experimental analysis performed on both a set of standard problem instances and new large problem instances demonstrate that accurate solutions are computed by the proposed techniques, especially for the distributed subpopulation version of the hybrid evolutionary algorithm. The comparative experimental evaluation demonstrates that the parallel evolutionary algorithms are able to outperform in reduced execution times the results computed using one of the best well-known deterministic techniques for the problem, in particular when solving instances with tight deadline constraints.

Parallel genetic algorithm based automatic path planning for crane lifting in complex environments

Abstract Heavy lifting is a common and important task in industrial plants. It is conducted frequently during the time of plant construction, maintenance shutdown and new equipment installation. To find a safe and cost effective way of lifting, a team works for weeks or even months doing site investigation, planning and evaluations. This paper considers the lifting path planning problem for terrain cranes in complex environments. The lifting path planning problem takes inputs such as the plant environment, crane mechanical data, crane position, start and end lifting configurations to generate the optimal lifting path by evaluating costs and safety risks. We formulate the crane lifting path planning as a multi-objective nonlinear integer optimization problem with implicit constraints. It aims to optimize the energy cost, time cost and human operation conformity of the lifting path under constraints of collision avoidance and operational limitations. To solve the optimization problem, we design a Master–Slave Parallel Genetic Algorithm and implement the algorithm on Graphics Processing Units using CUDA programming. In order to handle complex plants, we propose a collision detection strategy using hybrid configuration spaces based on an image-based collision detection algorithm. The results show that the method can efficiently generate high quality lifting paths in complex environments.

Aeroelastic tailoring of composite sandwich panel with lamination parameters

For sandwich panel with a thick core, stiffness matrix of composite sandwich panel can be formulated as function of membrane lamination parameters of the composite facesheet. In this paper, the flutter speed of a cantilevered sandwich panel is studied using lamination parameters of composite facesheet, with influence of transverse shear on sandwich panel considered. The results show that the flutter mode switch is not only related to the bending stiffness change, and also the transverse shear effect can result in switch of flutter mode. A minimum weight design of composite sandwich panel is conducted by using master-slave parallel real-coded genetic algorithm, with lamination parameters of facesheet and transverse shear modulus of foam core as design variables. The constraints considered include the flutter and divergence velocity constraints, the bound constraints of thicknesses of the facesheet and foam core. The feasible region constraints of membrane lamination parameters are also considered.

Multi-Objective Optimal Power Flow Considering Transient Stability Based on Parallel NSGA-II

Stability is an important constraint in power system operation and the transient stability constrained optimal power flow (OPF) has always received considerable attention in recent years. In this paper, the defects of the existing models and algorithms around this topic are firstly analyzed, on the basis of which, a multi-objective optimization method is proposed. The basic idea of the proposed method is to model transient stability as an objective function rather than an inequality constraint and consider classic transient stability constrained OPF (TSCOPF) as a tradeoff procedure using Pareto ideology. Second, a master-slave parallel elitist non-dominated sorting genetic algorithm II is used to solve the proposed multi-objective optimization problem, the parallel algorithm shows an excellent acceleration effect and provides a set of Pareto optimal solutions for decision makers to select. An innovative weight assigning technique based on fuzzy membership variance is also introduced for a more scientific and objective optimal solution decision. Case study results demonstrate the proposed multi-objective method has many advantages, compared with traditional TSCOPF methods.

Parallel relative radiometric normalisation for remote sensing image mosaics

Relative radiometric normalisation (RRN) is a vital step to achieve radiometric consistency among remote sensing images. Geo-analysis over large areas often involves mosaicking massive remote sensing images. Hence RRN becomes a data-intensive and computing-intensive task. This study implements a parallel RNN method based on the iteratively re-weighted multivariate alteration detection (IR-MAD) transformation and orthogonal regression. To parallelise the method of IR-MAD and orthogonal regression, there are two key problems: the normalisation path determination and the task dependence on normalisation coefficients calculation. In this paper, the reference image and normalisation paths are determined based on the shortest distance algorithm to reduce normalisation error. Formulas of orthogonal regression are acquired considering the effect of the normalisation path to reduce the task dependence on the calculation of coefficients. A master-slave parallel mode is proposed to implement the parallel method, and a task queue and a process queue are used for task scheduling. Experiments show that the parallel RRN method provides good normalisation results and favourable parallel speed-up, efficiency and scalability, which indicate that the parallel method can handle large volumes of remote sensing images efficiently.