Search

Due to the increasing dependence on mobile devices worldwide, web designers are starting to design websites that are suitable for mobile devices such as smartphones and tablets. The aim of this paper is to find out which structures of interface the mobile commerce websites should implement to meet the user’s viewing visual behaviors. This is achieved through the use of eye tracking technology to record and analyze the content of online booking websites designed by responsive web design rules and to assess whether there was a difference in the location of the attention areas when the webpage presents different object arrangements. The paper finds that there are no significant differences in the number of fixations at the location of the menu icon, and the gaze sequences are also the same. However, participants have significant differences in fixation time on the navigation menu lists with different presentation methods, whereas the gaze sequences are the same; on the other hand, there are no significant differences in the total gaze time when objects are arranged in different positions, but the gaze sequences on the regions of interest are different when the object arrangements are changed.

Autonomic computing was the term coined by IBM in 2001. The term autonomic computing was used to define the self-adaptable nature of the human body. According to IBM, the same self-adaptable feature was the need to be incorporated in the software systems. Autonomic computing is the combination of few self-capabilities such as self-configuration, self-healing, self-optimization, self-protection, self-awareness, etc. So, autonomic computing approach was then used to develop autonomic software systems. This approach makes the computing systems self-adaptable and self-decision-making support systems for various activities. It also helps to reduce the human intervention in the software management process. Though, the implementation of autonomic self-capabilities may increase the software complexity, which further requires human intervention for the software maintenance-related specific tasks. Still, IT industries are approaching to develop autonomic features in their existing architecture or developing new self-adaptable software systems. Autonomic computing has its importance for providing a bridge for handling and managing the run-time computation-based issues/exceptions of the software. So, the discussion of this solution has become a necessity for making the vision of autonomic decision making more clear and understandable for researchers and developers for the improvement in an autonomic area. The paper provides an insight vision of the autonomic decision-making concept and its importance for the various purposes such as intrusion detection, cloud-based data security, wireless sensor network, Internet of Things, Big Data and many other areas where management cannot be handled by a human in real time. To assess the degree of autonomic feature, there is another term used which is known as autonomicity. The paper also discusses some solutions suggested and implemented by different researchers during their studies for estimating the system’s autonomicity level. These solutions will help in comparing different autonomic applications based on the autonomic features implemented in each application. This paper is an attempt to provide better understandability in the autonomic computational field.

Cooperative multiple task assignment problem is an essential issue in the collaboration of multiple unmanned aerial vehicles (UAVs). Consensus-based bundle algorithm (CBBA) is a decentralized task assignment method that only considers homogeneous agents and independent tasks. Thus, we develop an extended CBBA with task coupling constraints (CBBA-TCC) in this paper to solve the multi-task assignment problem with task coupling constraints in the heterogeneous multi-UAV system. CBBA is a two-stage iteration algorithm with inner and outer consensus stages. The inner consensus stage is designed as a modified version of CBBA in this paper. A Can-do list is firstly raised at the beginning of bundle construction phase on each agent to record the tasks that can be performed by this agent without violating the task precedence constraints. Hence, at the inner consensus stage, each agent will only bid on the Can-do list. Then, we adopt a task performing time list for each agent to store the performing times of its assigned tasks. With associate consensus strategy of task performing time list at the conflict resolution phase, the precedence constraint of coupled tasks can be guaranteed. After reaching inner consensus, the outer consensus stage introduces an insert-position feasibility index to determine whether the assigned tasks satisfy the coupling constraints and resolve the constraint violation conflicts. Through the iterations of inner and outer consensus stages, CBBA will reach global consensus and obtain conflict-free task assignment results. Numerical simulations demonstrate the feasibility and reliability of CBBA in various search and rescue scenarios.

This paper presents a new opportunistic scheduling and resource consolidation system based on an economic model related to different service level agreements (SLAs) classes. The goal is to address the problem of companies that manage a private infrastructure of machines, i.e., a cloud platform and would like to optimize the scheduling of several requests submitted online by users. For the sake of simplicity of the presentation, the proposed economic model has two SLAs classes (qualitative and quantitative) with three Quality of Service for each SLA class (Premium, Advanced and Best effort). The consequence of this choice as well as the need to serve requests as they come have an impact on the algorithmic ways to consolidate an infrastructure. Indeed, our system proposes a new allocation heuristic that adapts the number of active machines in the cloud according to the global resources usage of all machines inside the infrastructure. This heuristic can be examined as a consolidation heuristic, based on the idea that the system can make reasonable choices, based on the SLAs, for the placement and the allocation of resources for each request. Experimentation with our system is conducted on Prezi (Web workload) and Google Cloud Data (HPC-oriented workload) traces, and they demonstrate the potential of our approach under different scenarios. From a methodological point of view, we propose a general framework which is limited in scope, for the sake of simplicity in reading the paper, with a small number of SLAs, but the idea can be extended to many more SLAs and performance metrics. In this way, the user or the provider operating the cloud have more latitude, thanks to our multi-criteria approach, to control the workload without a sacrifice on performance.

Our paper aims at learning from imbalance data based on ensemble learning. At the stage, the main solution is to combine under-sampling, oversampling or cost sensitivity learning with ensemble learning. However, these feature space-based methods fail to reflect the transformation of distribution and are usually accompanied with high computational complexity and risk of overfitting. In this paper, we propose a dynamic cluster algorithm based on coefficient of variation (or entropy), which learns the local spatial distribution of data and hierarchically clusters the majority. This algorithm has low complexity and can dynamically adjust the cluster according to the iteration of AdaBoost, adaptively synchronized with changes caused by sample weight changes. Then, we design an index to measure the importance of each cluster. Based on this index, a dynamic sampling algorithm based on maximum weight is proposed. The effectiveness of the sampling algorithm is proved by visual experiments. Finally, we propose a cost-sensitive algorithm based on Bagging, and combine it with the dynamic sampling algorithm to propose a multi-fusion imbalanced ensemble learning algorithm. In experimental research, our algorithms have been validated on three artificial datasets, 22 KEEL datasets and two gene expression cancer datasets, and have shown ideal or better performance than SOTA in terms of AUC, indicating that our algorithms are not only effective imbalance algorithms, but also provide potential for building a reliable biological cyber-physical system.

Multiple sequence alignment (MSA) is a core problem in many applications. Various optimization algorithms such as genetic algorithm and particle swarm optimization (PSO) have been used to solve this problem, where all of them are adapted to work in the bioinformatics domain. This paper defines the MSA problem, suggests a novel MSA algorithm called ‘NestMSA’ and evaluates it in two domains: Web data extraction and removing different URLs with similar text (DUST). The suggested algorithm is inspired by the PSO optimization algorithm. It is not a generalization of a two-sequence alignment algorithm as it processes all the sequences at the same time. Therefore, it looks globally at the same time on all sequences. Different from other PSO-based alignment algorithms, swarm particles in the proposed NestMSA algorithm are nested inside the sequences and communicated together to align them. Therefore, global maximum is guaranteed in our algorithm. Furthermore, this work suggests a new objective function which both maximizes the number of matched characters and minimizes the number of gaps inserted in the sequences. The running time complexity and the efficiency of NestMSA are addressed in this paper. The experiments show an encouraging result as it outperforms the two approaches DCA and TEX in the Web data extraction domain (95% and 96% of recall and precision, respectively). Furthermore, it gives a high-performance result in the DUST domain (95%, 93% and 92% of recall, precision and SPS score, respectively).

The power system operation, stability, and reliability are strongly dependent on the network protection system performance. On the other hand, the penetration of the renewable energy power plants into power grids drastically affects power networks and smart grids’ stability and reliability because of their fast dynamics in comparison to conventional power generators. Hence, power networks are considered as a complex cyber–physical energy system with vast interactions and communications among consumers and generators. In this condition, a reliable monitoring and control system can augment the safety, security, and operation of such an intricate system. Suitable time coordination between the operations of power grids protection relays is one of the most challenging issues to protect the power system properly. In this paper, a reliable supervisor system (RSS) with fast and precise control ability is proposed to improve power systems and smart grids protection system operations. The device utilizes a safe robust communication way which provides a condition for the available relays on two sides of a line to share their data. For this reason, the proposed device is able to diagnose the fault on a short transmission line quickly upon the event occurrence. In other words, fast communication between two sides of a short transmission line helps the protection system to work faster in the presence of a fault. In addition, quick handshaking and data sharing of the network different points provided by the RSS enable grid circuit breakers to disconnect a fault from the system as fast as possible. The proposed system, then, can easily find a defect location and disconnect a faulty line from the grid. According to the experimental results provided in this paper, the RSS can effectively corroborate the network stability.

The resources in the edge cloud are numerous and complex, and elastic scaling services can make efficient use of these resources. However, the elastic scaling services need to suspend the user’s application tasks forcibly when carrying out resource redistribution, which brings a poor sense of experience to the user. In view of the above problems, a dynamic elastic scaling model based on load prediction is proposed, which improves resource utilization and reduces scaling costs without affecting user experience. The model is divided into two parts. In terms of load prediction, on the one hand, according to the historical features and current trends of the load, the load prediction model based on the improved cloud model is used to predict the load demand at the next moment. On the other hand, the correlation between CPU and memory is considered. In terms of elastic scaling, integer programming algorithm is proposed to expand and release the corresponding resources with the minimum cost of horizontal scaling (HS) and vertical scaling (VS). In order to verify the superiority of elastic scaling model based on load prediction, corresponding comparative experiments are conducted, which show that the proposed model can improve the accuracy of load prediction and resource utilization with low scaling costs. Especially, the cost of elastic scaling proposed by this paper is lower than horizontal and vertical scaling. Compared with HS, the elastic scaling method proposed in this paper reduces the cost by 14%. Compared with VS, this method reduces the cost by 11%.

Due to the development of the Internet and the smart end systems such as smartphones, portable laptop, and other smart mobile devices, as well as the emergence of concepts such as cloud computing, social networks, and Internet of Things, 4G and 5G have changed network requirements. A software-defined network (SDN) is a new architecture to support new network requirements. This architecture is composed of three layers, data, control, and application plane. A lot of papers have been published in well-known journals to improve network performance, so we classify the solutions for performance improvement based on three SDN layers and three SDN-based spheres which are (a) SDN-data center, (b) wireless networks, and (c) software-defined wide area network. This paper surveys the articles which have been published by Springer, IEEE, Elsevier, and ACM. We extract the performance improvement solutions and resources which are required for performance control and the metrics to evaluate the network performance. This article can help network enthusiasts to better understand, investigate, and improve the performance of SDN-based networks.

Spark is one of the most widely used systems for the distributed processing of big data. Its performance bottlenecks are mainly due to the network I/O, disk I/O, and garbage collection. Previous studies quantitatively analyzed the performance impact of these bottlenecks but did not focus on iterative algorithms. In an iterative algorithm, garbage collection has more performance impact than other workloads because the algorithm repeatedly loads and deletes data in the main memory through multiple iterations. Spark provides three caching mechanisms which are “disk cache,” “memory cache,” and “no cache” to keep the unchanged data across iterations. In this paper, we provide an in-depth experimental analysis of the effect of garbage collection on the overall performance depending on the caching mechanisms of Spark with various combinations of algorithms and datasets. The experimental results show that garbage collection accounts for 16–47% of the total elapsed time of running iterative algorithms on Spark and that the memory cache is no less advantageous in terms of garbage collection than the disk cache. We expect the results of this paper to serve as a guide for the tuning of garbage collection in the running of iterative algorithms on Spark.

Clustering or partition is a fundamental work for graph or network. Detecting communities is a typical clustering, which divides a network into several parts according to the modularity. Community detection is a critical challenge for designing scalable, adaptive and survivable trust management protocol for a community of interest-based social IoT system. Most of the existed methods on community detection suffer from a common issue that the number of communities should be prior decided. This urges us to estimate the number of communities from the data by some way. This paper concurrently considers eliminating the number of communities and detecting communities based on block diagonal dominace adjacency matrix. To construct a block diagonal dominance adjacency matrix for the input network, it first reorders the node number by the breadth-first search algorithm. For the block diagonal dominance adjacency matrix, this paper shows that the numbers of nodes in a community should be continuous adjacent. And thus, it only needs insert some breakpoints in node number sequence to decide the number of communities and the nodes in every community. In addition, a dynamic programming algorithm is designed to achieve an optimal community detection result. Experimental results on a number of real-world networks show the effectiveness of the dynamic programming approach on the community detection problem.

The ultimate goal of modern cryptography is to protect the information resource and make it absolutely unbreakable and beyond compromise. However, throughout the history of cryptography, thousands of cryptosystems emerged and believed to be invincible and yet attackers were able to break and compromise their security. The main objective of this paper is to design a robust cryptosystem that will be suitable to be implemented in Internet of Things. The proposed cryptosystem is based on algebraic geometric curves, more specifically on Hermitian curves. The new cryptosystem design is called Hermitian-based cryptosystem (HBC). During the development of the HBC design, Kerckhoffs’s desideratum was the main guidance principle, which has been satisfied by choosing the Hermitian curves as the core of the proposed design. The proposed HBC inherits all the advantageous characteristics of Hermitian curve which are large number of points that satisfy the curve and high genus curves. The aforementioned characteristics play a crucial role in generating a large size encryption key for HBC and determine the block size of plaintext. Due to the fact that HBC used algebraic geometric codes over Hermitian curve, it has the ability to perform error correction in addition to data encryption. The error correction is another advantage of HBC compared with many existing cryptosystems such as McEliece cryptosystem. The number of errors that can be corrected by HBC is larger (high data rate) than other algebraic geometric codes such as elliptic and hyperelliptic curves. It also uses non-binary representation which increases its attack resistance. In this paper, the proposed HBC has been mathematically compared with elliptic curve cryptosystem. The results show that HBC has many advantages over the elliptic curves in terms of number of points and genus of the curve.

The increasing social problems on population, resources and environment enable the interaction between nature and humanity to become one of the most active research fields in the world. In this paper, we propose a novel framework of human–land sustainable computational system, which fully advances the progress of the development of our society utilizing the cloud computing and big data analysis technologies. Particularly, the study on quality of land management has attracted much attention. With the proposed framework, multi-user multi-cloud environment (MUMCE) is firstly presented, and evaluation of land quality is regarded as various services, such as soil acidity and alkalinity, soil thickness, soil texture, smoothness and field layout. Then, this paper formulates the problem of formal concept analysis-based multi-cloud composition recommendation with regard to multiple users. To address this problem, this paper first adopts the collaborative filtering to obtain the services request of the target user, then the service–provider concept lattices are constructed, and finally the best multi-cloud composition is selected and further recommended for the target user. Meanwhile, the corresponding algorithm is also devised. A case study is conducted for evaluating the feasibility and effectiveness of the proposed approach.

Modern computational platforms are characterized by the heterogeneity of their processing elements. Additionally, there are many algorithms which can be structured as a set of procedures or tasks with different computational cost. Balancing the computational load among the available processing elements is one of the main keys for the optimal exploitation of such heterogeneous platforms. When the processing time of any procedure executed on any of the available processing elements is known, this workload-balancing problem can be modeled as the well-known scheduling on unrelated parallel machines problem. Solving this type of problems is a big challenge due to the high heterogeneity on both, the tasks and the machines. In this paper, the balancing problem has been formally defined as a global optimization problem which minimizes the makespan (parallel runtime) and a heuristic based on a Genetic Algorithm, called Genetic Scheduler (GenS), has been developed to solve it. In order to analyze the behavior of GenS for several heterogeneous clusters, an example taken from the field of statistical mechanics has been considered as a case study: an active microrheology model. Given this type of problem and a heterogeneous cluster, we seek to minimize the total runtime to extend and analyze in depth the case of study. In such context, a task consists of the simulation of a tracer particle pulled into a cubic box with smaller bath particles. The computational load depends on the total number of the bath particles. Moreover, GenS has been compared to other dynamic and static scheduling approaches. The experimental results of such a comparison show that GenS outperforms the rest of the tested alternatives achieving a better distribution of the computational workload on a heterogeneous cluster. So, the scheduling strategy developed in this paper is of potential interest for any application which requires the execution of many tasks of different duration (a priori known) on a heterogeneous cluster.

This paper presents an approach for tolerant control and compensation of cyber attacks on the inputs and outputs of a cyber-physical system of rotary gantry type. The proposed control schemes are designed based on classic–intelligent control strategies for trajectory tracking and vibration control of a networked control system, which are developed for tip angular position control, while the system is prone to cyber attacks. The malicious attacks are assumed to be of denial of service (DoS) kind and cause packet loss with high probability in the two signals; control input and sensor output. In this paper, several classic and intelligent control strategies are studied in terms of robustness and effectiveness to attacks. Based on the results, a new hybrid control scheme is designed using linear quadratic regulation, sliding mode control, and artificial radial basis function neural network to alleviate the effect of DoS attacks and maintain the performance of the cyber-rotary gantry system in tracking applications. The neural network controller is trained during the control process. Its learning algorithm is based on the minimization of a cost function which contains the sliding surface. The hybrid control system is analyzed from the stability perspective. Moreover, the efficiency of the proposed scheme is validated by simulation on MATLAB Simulink platform.

Dynamic reconfiguration provides flexibility in the design and management of reconfigurable computing (RC) systems such that numerous applications would be mapped into limited resources simultaneously. As the mapping is a computationally intensive procedure in application compilation, a low-complex method is needed strongly for RC applications. In this paper, we propose a link-elimination partitioning approach for application graphs to reduce computations and reach an optimal solution faster as well. The link-elimination preprocessing step is performed by investigating the standard deviation of weights and removing lightweight links from the partitioning procedure. Based on the Laplacian matrix, a formulation method for detecting high-degree nodes as partition seeds has been generated. Moreover, a distance model for the region of implementation in resource graph has been introduced in this paper. In order to select among any rectangular shape of the resource graph, an average distance factor has been defined analytically. It has been proved that partitions with more connectivity must be implemented in a square-formed shape. Extensive experiments with random and benchmark graphs have been carried out to compare the proposed partitioning approach with the previous methods, and the results manifested that for fixed searching iterations, quality of solutions and time overhead have been improved 22% and 59%, respectively.

Since scientific workflow scheduling becomes a major energy contributor in clouds, much attention has been paid to reduce the energy consumed by workflows. This paper considers a multi-objective workflow scheduling problem with the budget constraint. Most existing works of budget-constrained workflow scheduling cannot always satisfy the budget constraint and guarantee the feasibility of solutions. Instead, they discuss the success rate in the experiments. Only a few works can always figure out feasible solutions. These methods work, but they are too complicated. In workflow scheduling, it has been a trend to consider more than one objective. However, the weight selection is usually ignored in these works. The inappropriate weights will reduce the quality of solutions. In this paper, we propose an energy-aware multi-objective reinforcement learning (EnMORL) algorithm. We design a much simpler method to ensure the feasibility of solutions. This method is based on the remaining cheapest budget. Reinforcement learning based on the Chebyshev scalarization function is a new framework, which is effective in solving the weight selection problem. Therefore, we design EnMORL based on it. Our goal is to minimize the makespan and energy consumption of the workflow. Finally, we compare EnMORL with two state-of-the-art multi-objective meta-heuristics in the case of four different workflows. The results show that our proposed EnMORL outperforms these existing methods.

More than 50 million journal papers will have been published by the end of 2019 with 2 million more journal papers published every year. The number of conference papers is even higher, and millions of other types of scientific research are added to the knowledge base every year. Scientific databases such as Web of Science, Scopus, and PubMed index millions of scientific papers and Google Scholar indexes a huge amount of scientific knowledge across diverse domains. However, current systems provide long lists of results when users attempt to find relevant papers, leaving them with little choice other than manually skimming through the lists. This article surveys different techniques used to identify relevant research papers by knowledge-based organizations. We categorized current literature content as content, metadata, collaborative filtering, and citation based techniques and identified the strengths and limitation for each approach. Further, we evaluated the published techniques and research-based products used to identify relevant documents and identified the strengths and limitations of each approach. This research will greatly help to understand current state-of-the-art techniques internal workings for finding relevant papers, understand the relevant strengths and limitations, and explore previously proposed techniques targeting this area.

Multi-robot task allocation problem is an important research field in multi-robot systems. The multi-robot task allocation problem combined with affective computing is an interesting frontier problem. This paper proposes a distributed affective robot pursuit task allocation algorithm based on self-awareness, which combines cognitive intelligence with emotional intelligence to establish self-awareness for affective robots. This paper defines the cognitive mechanism, environmental detection mechanism and self-decision mechanism of affective robots, so that the self-aware affective robots can make decisions autonomously by grasping the information of itself and the environment. The experimental results show better efficiency of this task allocation algorithm at each task scale, and the allocation efficiency is gradually improved as the number of robot increases.

Appropriate irrigation planning with reference evapotranspiration (ETo) plays an important role in managing the water resources for agriculture. This paper presents machine learning models for predicting ETo and thereby aiding in irrigation planning. Daily weather data which include minimum and maximum temperature, relative humidity, solar radiation, soil temperature and wind speed were used to forecast ETo. The datasets were trained, validated and tested using multiple linear regression, random forest and gradient boost regression (GBR) algorithms. The results of the machine learning algorithms with and without the proposed preprocess techniques were compared. Performance evaluations of these models were done by comparing the root mean square error, mean absolute error and coefficient of determination. It was found that the preprocessed GBR model edges past the other two models in predicting the ETo. The paper also discusses the mechanism of irrigation planning with the help of ETo, crop evapotranspiration (ETc) and irrigation system efficiency.

In this paper, a motion video sequence extraction algorithm based on cumulative frame differential clustering is proposed for the same scene. Firstly, the difference results of the two frames are clustered, and then median filtering technology and morphological method are used to extract and process the clustered results for obtaining moving foreground objects. During the process of video object image extraction, the color space clustering algorithm and filtering algorithm are introduced to reduce color noise in image and improve the accuracy of target object location and extraction efficiency. Finally, the framework of extraction system are designed and implemented. The high performance and accuracy of the proposed method are proved by quantitative analysis and extraction results comparison. The research in this paper has theoretical and practical value for promoting the development and application of key frame extraction technology and content-based video retrieval. The experimental results show that the algorithm has better performance, in terms of low computational complexity, real-time performance and better segmentation results.

Energy recharging has received much attention now because of the development of wireless recharging technology. This paper proposes a greedy constructing tree algorithm for shortest path in perpetual wireless recharging wireless sensor networks (GCTA) which enables the recharger to work while moving through the shortest path. The algorithm can construct the shorter path which passes all the sensors only once and guarantees that each sensor will be fully recharged. It computes the path in advance with lower complexity. And it also uses the graph theory, the tree theory, the Markov chain theorem and the powerful computing function of a computer. The contributions of the paper are itemized as follows: guaranteeing all the sensors to be fully recharged only once and reducing the length of the recharging path through constructing the greedy tree. In the experiment, this algorithm can construct the shorter path. So it is novel, general and can be applied widely.

Intrusion detection technology plays an important role in ensuring information security. This paper briefly describes the intrusion detection technology and its development history. Based on the analysis of power information network structure and its security partition, this paper proposes a power information network intrusion detection framework for the intrusion attack problem of power information network and elaborates the implementation of each module. The association rule analysis algorithm and the association relationship between network data stream features can effectively detect the intrusion behaviour in the power information network. Experiments show that the intrusion detection system can effectively detect the intrusion attacks in the power information network and effectively protect the power information.

Scientific applications are some of the most computationally demanding software pieces. Their core is usually a set of linear algebra operations, which may represent a significant part of the overall run-time of the application. BLAS libraries aim to solve this problem by exposing a set of highly optimized, reusable routines. There are several implementations specifically tuned for different types of computing platforms, including coprocessors. Some examples include the one bundled with the Intel MKL library, which targets Intel CPUs or Xeon Phi coprocessors, or the cuBLAS library, which is specifically designed for NVIDIA GPUs. Nowadays, computing nodes in many supercomputing clusters include one or more different coprocessor types. To fully exploit these platforms might require programs that can adapt at run-time to the chosen device type, hardwiring in the program the code needed to use a different library for each device type that can be selected. This also forces the programmer to deal with different interface particularities and mechanisms to manage the memory transfers of the data structures used as parameters. This paper presents a unified, performance-oriented, and portable interface for BLAS. This interface has been integrated into a heterogeneous programming model (Controllers) which supports groups of CPU cores, Xeon Phi accelerators, or NVIDIA GPUs in a transparent way. The contribution of this paper includes: An abstraction layer to hide programming differences between diverse BLAS libraries; new types of kernel classes to support the context manipulation of different external BLAS libraries; a new kernel selection policy that considers both programmer kernels and different external libraries; a complete new Controller library interface for the whole collection of BLAS routines. This proposal enables the creation of BLAS-based portable codes that can execute on top of different types of accelerators by changing a single initialization parameter. Our software internally exploits different preexisting and widely known BLAS library implementations, such as cuBLAS, MAGMA, or the one found in Intel MKL. It transparently uses the most appropriate library for the selected device. Our experimental results show that our abstraction does not introduce significant performance penalties, while achieving the desired portability.

The paradigm of computing has shifted from computation-centric to communication-centric designs. Network-on-chip has emerged as an alternative interconnect mechanism for future multi-core designs. Transistor integration is approaching its limit, and this increases the susceptibility of the interconnects toward failures. Research efforts are directed toward improving the fault tolerance of these interconnects. Fault-tolerant routing such as segment-based and up*/down* are static by nature and require reconfiguration to circumvent failures. Failures may disrupt the connectivity of the network, and new routing instance needs to be configured in case old routing instance is unable to offer full connectivity. In this paper, we identify nodes affected by the failures and propose an extended scalable routing reconfiguration, called S2DIO. It performs reconfiguration of affected nodes by taking $$N^2$$ cycles for $$N \times N$$ mesh network, whereas state-of-the-art (ARIADNE) consumes $$N^4$$ cycles. Instead of routing tables, we employ logic-based routing and achieve significant improvements, i.e., $$30.7\%$$ in terms of area and $$29\%$$ in terms of power overhead for a $$16\times 16$$ mesh router. A novel algorithm for computation of new logic-based routing bits is also proposed in this paper. Our reconfiguration (S2DIO) improves average flit latency up to $$32\%$$ and throughput up to 19% for single-link failure in $$8\times 8$$ 2D mesh network-on-chip.

Since Osman Abul et al. first proposed the k-anonymity-based privacy protection for trajectory data, researchers have proposed a variety of trajectory privacy-preserving methods. These methods mainly adopt a static anonymity algorithm, which only focusing on the trajectories in a specific time span, directly anonymizes data and publishes them without considering dynamic nature of trajectory data as the new time slice arriving. Furthermore, due to its correlation with time and position, the trajectory data is produced in large scale and many sensitive attributes; the traditional k-anonymity-based privacy-preserving models need to recalculate the last released trajectory data, which will increase the computing cost and reduce the availability of the released trajectories, are not fit for privacy protection in large-scale trajectory data. Therefore, this paper presents a method to dynamically publish the large-scale vehicle trajectory data with privacy protection under $$(k,\delta )$$ security constraints. According to the spatial and temporal characteristics of vehicle trajectory data, this paper first proposes a method to partition the trajectory data for storage and computation. We choose the sample point $$(x_{i},y_{i})$$ at time $$t_{i}$$ as partition points and store the partitions of the trajectory data according to the time sequence and location of the running vehicle. This results in the efficient trajectory scanning, clustering and privacy protection. We use $$(x_{i},y_{i},t_{m}-t_{n})$$ to represent the identifier of trajectory data to publish, use the generalize function to cluster trajectory data under the $$(k,\delta )$$ security constraints. Through this way, we can effectively process the trajectory in every data partition as time goes on and need not to recalculate the released trajectories, effectively reduce the computing cost. Through experiments on real trajectory data and Oldenburg trajectory data, confirming the data partitioning method in privacy-preserving large-scale trajectory data publishing under the security constraint of $$(k,\delta )$$, and the l-diversity. By the experimental comparison, our method maintains a least level of computing cost and higher data availability.

Present heterogeneous architectures interconnect nodes including multiple multi-core microprocessors and accelerators that allow different strategies to accelerate the applications and optimize their energy consumption according to the specific power-performance trade-offs. In this paper, a multilevel parallel procedure is proposed to take advantage of all nodes of a heterogeneous CPU–GPU cluster. Two more alternatives have been implemented, and experimentally compared and analyzed from both running time and energy consumption. Although the paper considers an evolutionary master–worker algorithm for feature selection in EEG classification, the conclusions from the experimental analysis here provided can be frequently applied, as many other useful bioinformatics and data mining applications show the same master–worker profile than the classification problem here considered. Our parallel approach allows to reduce the time by a factor of up to 83, with only about a 4.9% of energy consumed by the sequential procedure, in a cluster with 36 CPU cores and 43 GPU compute units.

This paper presents an internet-of-things (IoT)-cloud converged virtual machine (VM) system for IoT devices with restricted computing resources. The VM is a software processor that has many advantages in terms of software development, release, maintenance, etc., owing to its platform independence. However, in low-performance devices it has a significant disadvantage because its use is restricted by high execution overheads at the software interpretation level. This paper proposes a VM system for IoT devices that solves this problem while retaining the advantages of VM technology using a lightweight interpreter model and cloud-based computation offloading. The proposed interpreter solves the limited memory/performance problem when running VMs on low-performance devices using two-level instruction-to-native function matching techniques. Furthermore, by solving the low-performance issues of IoT devices using cloud-based offloading, the proposed IoT-cloud VM can run applications that require high-performance computing even when the target hardware system is a low-power IoT device.

In the domain of proteomics, an in-depth analysis of the 3D structure of a protein is of paramount importance for many biological studies and applications. At the secondary level, protein structure can be described in terms of motifs, recurrent patterns of smaller biological structures called secondary structure elements. In this paper, the focus is on the identification of geometrical motifs in different proteins using the Cross Motif Search (CMS) algorithm. Such task, due to the high computational cost of CMS with respect to traditional alignment algorithms, is very demanding, and thus, parallel processing is mandatory. In previous papers, CMS parallelization has been already studied from the HPC standpoint. Since cloud computing is emerging as an alternative to on-premise HPC systems, it is worthwhile examining the feasibility and possible advantages in terms of both performance and costs, of migrating to a cloud implementation. This paper is an extension of a preliminary work carried out on the cloud parallelization of CMS. The paper has two main contributions. First of all, an analytic model of the communication pattern of CMS is described, in order to get insights on the performance of the application when executed on a cloud infrastructure. Secondly, an optimized “location-aware” scheduling policy to assign workload to the application workers is introduced, in order to minimize internode communication in a cloud setting. Experiments are presented in order to validate the newly introduced scheduling policy and assess the performance of the cloud implementation of CMS. The results presented in this paper are general, in the sense that they can be applied to any other algorithm with a communication pattern similar to the one of the target applications.

In this paper, we propose speech/music pitch classification based on recurrent neural network (RNN) for monaural speech segregation from music interferences. The speech segregation methods in this paper exploit sub-band masking to construct segregation masks modulated by the estimated speech pitch. However, for speech signals mixed with music, speech pitch estimation becomes unreliable, as speech and music have similar harmonic structures. In order to remove the music interference effectively, we propose an RNN-based speech/music pitch classification. Our proposed method models the temporal trajectories of speech and music pitch values and determines an unknown continuous pitch sequence as belonging to either speech or music. Among various types of RNNs, we chose simple recurrent network, long short-term memory (LSTM), and bidirectional LSTM for pitch classification. The experimental results show that our proposed method significantly outperforms the baseline methods for speech–music mixtures without loss of segregation performance for speech-noise mixtures.

Localization is a process of finding the coordinates of the deployed sensor nodes in the sensing area. Localization in wireless sensor networks (WSNs) is important as it results in location-stamped communication. Applications of WSNs like forest fire detection, nuclear or biological attacks, locating survivors post-disasters, etc. require location-aware information for real time and accurate response. Generally, the sensor nodes are not equipped with global positioning system due to its inefficiency for indoor and underwater regions. Other drawbacks include cost and power consumption. Localization in 3D WSNs is a complex task. The complexity increases due to an additional dimension leading to larger neighboring nodes and increase in coverage area (i.e., area coverage of a sensor node changes from circular to spherical). A lot of approaches have been proposed to localize the sensor nodes in 2D and are now being seen in 3D aspect. Although good efforts have been made for research summarization and guidance in 2D localization, it lacked attention by the researchers in 3D scope. This paper provides a comprehensive survey of the algorithms designed for localizing the sensor nodes in terrestrial and underwater regions. These algorithms have been classified based on the nature of anchor nodes. The localization methods have been categorized as static anchor node-based or mobile anchor node-based and further range-free or range-based depending upon the number of anchor nodes availability and distance estimation technique, respectively. The limitations and challenges of the proposed approaches have also been discussed briefly. The paper also tabulates these methods on the basis of attributes like localization process, complexity, number of sensor nodes used, etc.

This paper proposes a method to reconstruct three-dimensional (3D) objects using real-time fusion and analysis of multiple sensor data. This paper attempts to create a realistic 3D visualization with which a remote pilot can intuitively control a remote unmanned robot by utilizing the characteristics of massive sensor data. The 3D reconstruction system proposed in this paper is comprised of 3D and two-dimensional (2D) data segmentation method, a 3D reconstruction method applied to each object, and a projective texture mapping method. Specifically, we propose applying both a 2D region extraction method and a 3D mesh modeling method to each object. The proposed schemes are implemented as a real-time application to verify real-time performance. This paper proves that 3D meshes can be modeled in real time by using the proposed method. The proposed method allows the remote control of a robot for real-time 3D rendering of remote scenes, which is essential for various tasks in areas that cannot be easily accessed by humans.

This paper proposes a smart machine-learning-based unmanned aerial vehicle (UAV) control system that optimizes UAV operations in real time. The purpose of this system is to increase the efficiency of UAVs that need to operate with limited resources. This can be accomplished by allowing the UAVs in flight to identify their current state and respond appropriately. The proposed system, which is developed based on “cloud robotics,” benefits from the powerful computational capabilities of cloud computing and can therefore calculate many types of information received from various sensors in real time to maximize the performance of the UAV control system. The system learn about normal situations when creating models. That is, preprocessing data that is correlated with a particular situation and modeling it with a “multivariate Gaussian distribution.” Once the model is created, the UAV can be used to analyze the current situation in real time during flight. Of course, it is possible to recognize the situation based on the traditional RULE or the latest LSTM. However, this is not an appropriate solution for UAV situations where irregularities are severe and unpredictable. In this paper, we succeeded in recognizing the UAV flight status in real time by the proposed method and succeeded in optimizing it by adjusting communication cycle based on a recognized situation. Based on the results of this study, we expect to be able to stabilize and optimize systems that are highly irregular and unpredictable. In other words, this system will be extended to learn about various situations and create a model. A reliable and efficient smart system can be designed by judging the situation comprehensively.

Recent trends in big data have shown that the amount of data continues to increase at an exponential rate. This trend has inspired many researchers over the past few years to explore new research direction of studies related to multiple areas of big data. The widespread popularity of big data processing platforms using MapReduce framework is the growing demand to further optimize their performance for various purposes. In particular, enhancing resources and jobs scheduling are becoming critical since they fundamentally determine whether the applications can achieve the performance goals in different use cases. Scheduling plays an important role in big data, mainly in reducing the execution time and cost of processing. This paper aims to survey the research undertaken in the field of scheduling in big data platforms. Moreover, this paper analyzed scheduling in MapReduce on two aspects: taxonomy and performance evaluation. The research progress in MapReduce scheduling algorithms is also discussed. The limitations of existing MapReduce scheduling algorithms and exploit future research opportunities are pointed out in the paper for easy identification by researchers. Our study can serve as the benchmark to expert researchers for proposing a novel MapReduce scheduling algorithm. However, for novice researchers, the study can be used as a starting point.

The hypercube is one of the most popular interconnection networks. Its network cost is $$O(n^2)$$ . In this paper, we propose a new hypercube variant, the divide-and-swap cube $$\textit{DSC}(n)\,(n=2^d,\,d\ge 1)$$ , which reduces the network cost to $$O(n \log n)$$ while maintaining the same number of nodes and the same asymptotic performances for fundamental algorithms such as the broadcasting. The new network has nice hierarchical properties. We first show that the diameter of $$\textit{DSC}(n)$$ is lower than or equal to $$\frac{5n}{4}-1$$ . However, unlike the hypercube of dimension n whose degree is n, the node degree of the network is $$\log n + 1$$ , resulting in a network cost of $$O(n \log n)$$ . We then examine the one-to-all and all-to-all broadcasting times of $$\textit{DSC}(n)$$ , based on the single-link-available and multiple-link-available models. We also present an upper bound on the bisection width of the $$\textit{DSC}(n)$$ and show that $$\textit{DSC}(n)$$ is Hamiltonian. Finally, we introduce the folded divide-and-swap cube, $$\textit{FDSC}(n)$$ , a variant of the $$\textit{DSC}(n)$$ and study its many properties including its hierarchical structure, routing algorithm, broadcasting algorithms, bisection width, and its Hamiltonicity. All the broadcasting algorithms presented in this paper are asymptotically optimal.

This paper presents the development of artificial neural network (ANN) model for predicting the dynamic across-wind response of tall building as per the provisions given in Indian Wind Code (IWC), IS 875 (Part-3):2015. IWC has given the procedure for estimating across-wind load response for limited aspect ratios of the building. For calculating the across-wind response not conforming to the ratios specified in the code, additional time-consuming and expensive studies are required. In order to minimise such tedious studies, an efficient computational technique is essential to determine wind response. ANN proves to be a promising alternative technique for analysing complex engineering problems. Thus, in this paper, using the best-fit model of ANN, predictions are made for the building configurations that are not included in IWC. The results are plotted as a series of charts in terms of base shear and the base bending moment. It is found that the predictions made using the trained and developed ANN models are in good agreement with the desired output.

This paper concentrates on the entropy estimation of time series. Two new algorithms are introduced: Fast Approximate Entropy and Fast Sample Entropy. Their main advantage is their lower time complexity. Examples considered in the paper include interesting experiments with real-world data obtained from IT4Innovations’ supercomputers Salomon and Anselm, as well as with data artificially created specifically to test the credibility of these new entropy analyzers.

The variation of the network topology directly indicates the knowledge network evolution, the principal composition of which is the structural evolution. Based on the principles of both the minimum coupling costs between knowledge subjects and the network average value threshold, this paper constructs the evolution model of a small-world knowledge network. Using MATLAB programming, this paper analyzes the characteristics of the evolution of knowledge network. It then examines the main parameter changes and the distribution of the network node degree as well as its probability distribution and then proceeds to investigate the variation of the subjects’ knowledge level as well as the variation speed. The simulation results show that with the structural evolution of the knowledge network, the connection between the network subjects has been improved after numerous rewiring. Compared with the initial state, the total knowledge stock of the knowledge network and the average knowledge stock of the subjects are significantly enhanced.

Wireless sensor networks (WSNs) are resource-constrained networks with limited battery power, computation abilities and communication capabilities at the sensor node level. A long-standing need in the area is to reduce energy consumption of the WSN nodes and to extend the network lifetime, along with improving the network reliability and the network quality. Energy-efficient routing protocols can improve the WSN performance, and the algorithms have been explored widely by the researchers. In the majority of the work, sensor nodes have been considered homogeneous in their configuration and capabilities. However, in practical scenarios and with the advent of an era of connected devices, called "Internet of Things," consideration of node heterogeneity in WSN has become inevitable. Nodes' heterogeneity can be exploited constructively in WSN routing algorithms. In the literature, many different heterogeneous WSN scenarios (e.g., energy, computation, link heterogeneities) have been considered and the routing algorithms have been proposed for performance improvement in such scenarios. This paper focuses on the routing concepts for diverse heterogeneous WSNs scenarios and covers the state of the art in the area. The clustering-based approaches for routing decisions are extensively covered under different heterogeneous WSN scenarios. Along with presenting a comprehensive review of the routing algorithms under different varieties of heterogeneous WSNs, the paper also discusses the effects and interdependencies of different heterogeneities in routing decisions and unveils new research directions in the area.

Most of the wireless sensor networks (WSNs) are equipped with battery-powered devices with limited processing/communication resources which necessitate the designed algorithms to reduce the computational burden on nodes as much as possible. This could be achieved using distributed algorithms in which the computations are distributed between all of the nodes. In WSNs, the sensors observe phenomena that could be common to some of them, in addition to the intra-signal correlation; therefore, the acquired signals by the sensors possess some inter-signal correlation. These joint structures could be exploited if the designed algorithms are cooperative. On this basis, in this paper a new distributed and cooperative signal recovery algorithm in the application of compressive sensing for WSNs is proposed. We consider a situation that the sensor nodes intend to recover signals that differ from one sensor to another, in the underdetermined systems of equations, while these signals have common sparsity profile. In fact, we introduce a general structure which can be applied to many optimization problems with different non-convex objective functions and different constraints. In this paper, this method will be used on three problems with different constraints and will result in three completely distributed methods. The simulation results show that the proposed method presents better performance compared to the other existing algorithms in terms of both recovery and convergence rate.

MERS-CoV is an airborne disease which spreads easily and has high death rate. To predict and prevent MERS-CoV, real-time analysis of user's health data and his/her geographic location are fundamental. Development of healthcare systems using cloud computing is emerging as an effective solution having benefits of better quality of service, reduced cost, scalability, and flexibility. In this paper, an effective cloud computing system is proposed which predicts MERS-CoV-infected patients using Bayesian belief network and provides geographic-based risk assessment to control its outbreak. The proposed system is tested on synthetic data generated for 0.2 million users. System provided high accuracy for classification and appropriate geographic-based risk assessment. The key point of this paper is the use of geographic positioning system to represent each MERS-CoV users on Google maps so that possibly infected users can be quarantined as early as possible. It will help uninfected citizens to avoid regional exposure and the government agencies to manage the problem more effectively.

A novel and fast nonlocal mean (NLM) image denoising method using a structure tensor for Gaussian noise removal is proposed in this paper to address the difficult problem of the high computing cost of similarity weights. This method uses visual features based on fuzzy theory to measure the similarity between image pixels. First, from grayscale values and structure information in image patches, the similarity index measure is constructed by using a fuzzy theory measure. Then, the structure tensor is used to eliminate a large number of pixels of low similarity by the similarity index of the visual feature, which can improve the operation speed and avoid the filter parameters. Finally, NLM with a structure tensor and the similarity measure of fuzzy metrics is used to denoise the image. Due to the improved method considering the characteristics of the image visual structure, the structural similarity of NLM is improved. Compared with other improved NLM methods, the improved method in this paper has higher performance in denoising and reduces the complexity of the similarity computation and the problem of parameter setting.

In cooperative cognitive radio networks (CCRNs), resource allocation can be viewed as a multi-objective optimization issue in terms of channel capacity as well as, among numerous others, the transmitted power, and the QoS limitations. Many researchers have been undertaken to overcome individual problems, not multi-objective problems. In this paper, we investigate multi-objective problems, such as energy consumption, queuing problems, priority levels of traffic classes, fairness, throughput, and user quality requirements. We propose a hybrid optimization algorithm for CCRNs (HCCRN), which enhances the resource allocation. The first contribution of this paper is to propose the load balance enhanced particle swarm optimization algorithm for energy-efficient cluster formation, which overcomes queuing problems. In the second contribution, we consider multiple factors as the input of a multi-factor differential evolution optimization algorithm for prioritizing the traffic levels. The third contribution is that the fair routing path is computed by a modified gravitational search algorithm that enhances resource allocation throughput. For testing purpose, the proposed HCCRN algorithm applied to IEEE 802.11 WLANs. Simulation results show that the users achieve required resources via the proposed HCCRN, thus providing energy efficiency, fairness, throughput, and QoS.

In view of the disturbance of the emergency signal preemption to the normal traffic flow, a scientific and reasonable signal transition optimization model was proposed in this paper, in which fair and efficiency are both considered. Since the emergency signal preemption may lead to longer vehicle queuing and greater vehicle delay at intersections, the difference of queue length and the vehicle delay were selected as the objectives taking into account the fair benefit, and a multi-objective signal recovery transition optimization model for emergency rescue was established based on the power function method. The queue length was calculated by the 95% queue length model of the SYNCHRO, and the average vehicle delay was calculated with HCM2000 delay model; furthermore, parameters’ weight in the model was calibrated with the variation coefficient method, and model was solved by genetic algorithm. Moreover, power function method was used to score the signal recovery transition schemes of emergency traffic. Finally, based on the survey data from the typical route network in the Lion Mountain area of Suzhou City, the model proposed in this study was compared with three classical smooth transition schemes (immediate transition scheme, two-cycle transition scheme and three-cycle transition scheme) at the scenario of evening peak. Simulation results showed that the model in this paper can reduce the vehicle delay and queue length compared with the above three smooth transition schemes; the average reduction ratio of queue length delay was 13.82%, and this number of queue length was 13.65%, from which we can conclude that the model proposed in this has better performance than other three classical transition schemes in emergency signal transition applicability.

An electromagnetic flow sensor applied to the flow measurement of iron ore pipeline transportation will incur a measurement error after being used for a period of time, and the error will increase continuously. In this paper, based on the weight function theory of electromagnetic flow sensors, the main cause of the error is proposed. Specifically, in the process of using the electromagnetic flow meter the insulation lining of the sensor is easily damaged because the measuring pipe of the sensor constantly collides with the solid phase of the fluid, leading to a change in the distribution of the weight function of the electromagnetic flow sensor, which, in turn, changes the contribution of the induced electromotive force generated by each medium in the pipe to the measuring electrode. This paper is based on the PDE computer numerical calculation method, and, according to the analysis of the weight function under different boundary conditions, the rule that the output signal of the electromagnetic flow sensor decreases with increasing insulation lining area is obtained. The computer numerical calculation method used in this paper can provide a new approach to the theoretical analysis of the weight function of electromagnetic flow sensors. Furthermore, the conclusion reached in this paper can provide a theoretical basis for the development and maintenance of electromagnetic flow sensors applied to the flow measurement of iron ore pipelines.

The load balancing algorithm of web server cluster in network education system is studied in this paper, and a new strategy based on server status and request information is proposed to reduce the response time. The mechanism of the algorithm is to classify the degree of influence of the client’s request content on the server, and set the corresponding weight. At the same time, it combines the server performance parameters and load conditions to allocate the request to the lightest load of the server. The overall framework structure of the hierarchy is proposed according to the MOOC education system, and the internal communication mechanism of the system and the monitoring program of the web server performance are studied in this paper. Experimental results show that the algorithm can fully utilize the existing software and hardware resources, effectively balance the load of the web cluster server, and fully utilize the service resources.

Our goal is to provide a novel method of representing 2D shapes, where each shape will be assigned a unique fingerprint - a computable approximation to a conformal map of the given shape to a canonical shape in 2D or 3D space (see page 22 for a few examples). In this paper, we make the first significant step in this program where we address the case of simply, and doubly-connected planar domains. We prove uniform convergence of our approximation scheme to the appropriate conformal mapping. Along the way, we affirm a conjecture raised by Ken Stephenson in the 90's which predicts that the Riemann mapping can be approximated by a sequence of electrical networks. In fact, we first treat a more general case. Consider a planar annulus, i.e., a bounded, two-connected, Jordan domain, endowed with a sequence of triangulations exhausting it. We construct a corresponding sequence of maps which converge uniformly on compact subsets of the domain, to a conformal homeomorphism onto the interior of a Euclidean annulus bounded by two concentric circles. The resolution of Stephenson's Conjecture then follows by a limiting argument. Our methods involve harmonic mappings and boundary value problems: discrete and analytic in the plane. The scheme we constructed is programmable., Comment: This is the (more or less) published paper that appeared in The Journal of Supercomputing, (74), 2018, 6333--6368

The DVM and SAPFOR systems were primary designed to simplify the development of parallel programs of scientific-technical calculations. DVM includes CDVMH and Fortran-DVMH programming languages which are the extensions of standard C and Fortran languages by parallelism specifications. The DVMH programs can be efficiently executed on heterogeneous computational clusters with accelerators. SAPFOR is a software development suit that can be used to produce a parallel version of a sequential program in a semiautomatic way, according to DVMH model of the parallel programming. The paper considers the application of DVM and SAPFOR systems at all stages of the program parallelization on the example of mapping of 3D elastic waves simulation method on high-performance heterogeneous clusters with GPUs. SAPFOR enables us to investigate original sequential program. Then, Fortran-DVMH language is used to exploit parallelism. Tools for functional and performance debugging, that are part of the DVM system, allow to resolve errors in the parallel program and to improve its performance. The paper also compares the DVMH-based program with a program obtained after manual parallelization using MPI programming technology.

Floating-point error is one of the important factors that limit the accuracy of calculation. Programs using floating-point numbers are prone to accuracy problems by rounding and catastrophic cancelation. Mathematical library is an essential part of computer system. Its implementation is a common genre of floating-point calculation. Implementing these functions through polynomial expansion has algorithmic error in addition to rounding and catastrophic cancelation. Furthermore, ‘concise’ mathematical functions are frequently developed for applications with high-performance requirements. These ‘concise’ functions pay more attention to performance rather than accuracy. As a result, serious accuracy problems might occur, which will further affect the execution of programs. In this paper, a heuristic approach is designed to detect the maximum error of mathematical functions. It is theoretically proved in Sect. 2 that the maximum error of the mathematical functions generally occurs in small intervals. This finding is further proved by several experiments. The heuristic approach is designed based on it. Firstly, the large-scale test is used to gain the error distribution of functions. After that, the intervals, where the maximum error may occur (called ‘tingle intervals’ in this paper), are derived from the error distribution. At last, the high coverage test over tingle intervals is used to detect the maximum error. We evaluate the algorithm by analyzing mainstream math libraries (MKL, Glibc and Sunway TaihuLight). Our experiments show the approach can correctly deal with over 96.2% of mathematical functions. The approach designed in this paper has been used in the analysis and optimization of Sunway TaihuLight Math Library.

This paper built a system containing the Distributed Crawling Module, the Database Module and the Analysis Module to collect a large number of objective data, clean the data and realize the business intelligence representation. The distributed crawling system includes the Distributed Crawling Module building from Hadoop, the Database Module by SQL Server and the Analysis Module by the SAS system. The first two modules support a distributed way to collect and convert non-structural data into structural data for the last module processing. Then, this paper extends research fields from theory to application about BR the credit rating of Latvia, Estonia, Slovakia, Turkey, Malaysia, Russia, Thailand and other countries is very high; Afghanistan, Ukraine, Laos, Iran, Arabia, Republic of Syria, Iraq, Burma, Republic of Yemen, East Timor and other countries with poor credit ratings. The conclusion is consistent with the domestic and overseas well-known rating agencies.