Computational Operations Research Articles

Overcoming the Curse of Dimensionality in the Numerical Approximation of Parabolic Partial Differential Equations with Gradient-Dependent Nonlinearities

Partial differential equations (PDEs) are a fundamental tool in the modeling of many real-world phenomena. In a number of such real-world phenomena the PDEs under consideration contain gradient-dependent nonlinearities and are high-dimensional. Such high-dimensional nonlinear PDEs can in nearly all cases not be solved explicitly, and it is one of the most challenging tasks in applied mathematics to solve high-dimensional nonlinear PDEs approximately. It is especially very challenging to design approximation algorithms for nonlinear PDEs for which one can rigorously prove that they do overcome the so-called curse of dimensionality in the sense that the number of computational operations of the approximation algorithm needed to achieve an approximation precision of size {varepsilon }> 0 grows at most polynomially in both the PDE dimension d in mathbb {N} and the reciprocal of the prescribed approximation accuracy {varepsilon }. In particular, to the best of our knowledge there exists no approximation algorithm in the scientific literature which has been proven to overcome the curse of dimensionality in the case of a class of nonlinear PDEs with general time horizons and gradient-dependent nonlinearities. It is the key contribution of this article to overcome this difficulty. More specifically, it is the key contribution of this article (i) to propose a new full-history recursive multilevel Picard approximation algorithm for high-dimensional nonlinear heat equations with general time horizons and gradient-dependent nonlinearities and (ii) to rigorously prove that this full-history recursive multilevel Picard approximation algorithm does indeed overcome the curse of dimensionality in the case of such nonlinear heat equations with gradient-dependent nonlinearities.

Image processing method based on aerospace monitoring

The paper considers a digital image processing method that provides efficient transmission, redistribution and storage of aerospace environmental monitoring videographic information. The requirement to increase the monitoring efficiency, if necessary, to carry out regular monitoring of changes in the image field of the observed object is especially relevant when describing spatially aligned sets of images obtained after multi-zone, multi-time and multi-polarization shooting. In many remote sensing applications, the processing efficiency reflects the degree of reduction in the volume of transmitted, described, analyzed and stored videographic information. The computational processing of digital images based on the coordinate survey scheme is presented. The method can be used to perform the procedure for decoding images in order to solve the problem of classifying objects of interest, analyzing images. Such characteristics of binary objects as a spatial boundary and a contour were used as deciphering signs. These features make it possible to describe the shape of an object, its geometric parameters and search for images with specific spatial structures. The processing method is realized by executing efficient algorithms for coding an image on a discrete grid by means of a chain code and spectral coding based on the fast discrete Hartley transform. The result of processing comes down to minimizing the number of such basic computational operations as multiplication, addition, transferring, as well as reducing the time and capacitive complexity of the program, and reducing the redundancy of the initial data. An example of an effective representation and description of a segmented image of an object is given. The proposed method makes it possible to expand the technical capabilities of more efficient transmission and processing of images for solving problems in the field of remote sensing.

Optimized DA-reconfigurable FIR filters for software defined radio channelizer applications

PurposeIn this paper, the authors present different methods for reconfigurable finite impulse response (RFIR) filter design. Distributed arithmetic (DA)-based reconfigurable FIR filter design is suitable for software-defined radio (SDR) applications. The main contribution of reconfiguration is reuse of registers, multipliers, adders and to optimize various parameters such as area, power dissipation, speed, throughput, latency and hardware utilizations of flip-flops and slices. Therefore, effective design of building blocks will be optimized for RFIR filter with all the above parameters.Design/methodology/approachThe modified, direct form register structure of FIR filter contributes the reuse concept and allows utilization of less number of registers and parallel computation operations. The disadvantage of DA and other conventional methods is delay increases proportionally with filter length. This is due to different partial products generated by adders. The usage of adder and multipliers in DA-FIR filter restricts the area and power dissipation because of their complexity of generation of sum and carry bits. The hardware implementation time of an adder can be reduced by parallel prefix adder (PPA) usage based on Ling equation. PPA uses shift-add multiplication, which is a repetitive process of addition, and this process is known as Bypass Zero feed multiplicand in direct multiplication, and the proposed technique optimizes area-power product efficiently. The modified DA (MDA)-based RFIR filter is designed for 64 taps filter length (N). The design is developed by using Verilog hardware description language and implemented on field-programmable gate array. Also, this design validates SDR channel equalizer.FindingsBoth RFIR and SDR are integrated as single system and implemented on Artix-7 development board of XC7A100tCSG324 and exploited the advantages in area-delay, power-speed products and energy efficiency. The theoretical and practical comparisons have been carried out, and the results are compared with existing DA-RFIR designs in terms of throughput, latency, area-delay, power-speed products and energy efficiency, which are improved by 14.5%, 23%, 6.5%, 34.2% and 21%, respectively.Originality/valueThe DA-based RFIR filter is validated using Chipscope Pro software tool on Artix-7 FPGA in Xilinx ISE design suite and compared constraint parameters with existing state-of-art results. It is also tested the filtering operation by applying the RFIR filter on Audio signals for removal of noisy signals and it is found that 95% of noise signals are filtered effectively.

(Invited) The Quantum Phase Interference in the Excited Triplet of Fullerene

High spin magnetic molecules are promising candidates for quantum information processing because their intrinsic multiplicity facilitates information storage and computational operations. However, due to the absence of suitable sublevel splittings, their susceptibility to environmental disturbances and limitation from the selection rule, the arbitrary control of the quantum state of a molecular electron multiplet has not been realized. Here we exploit the photoexcited triplet of C70 as a molecular electron spin qutrit with pulsed electron paramagnetic resonance. We prepared the system into 3-level superposition states characteristic of a qutrit and validated them by the tomography of their density matrices. To further elucidate the coherence of the operation and the nature of the system as a qutrit, we demonstrated the quantum phase interference in the superposition. The interference pattern is further interpreted as a map of possible evolution paths in the space of phase factors, representing the quantum nature of the 3-level system. Figure 1

Routing Protocol Based On Energy Efficient For Under Water Wireless Network

Underwater Wireless Sensor Networks is effective and intelligent utilization of energy for routing protocol in longer network lifetime. Energy Consumption and load balancing are the vital roles for a network lifetime. The use of load balancing in WSN is granted as the best resource of sink mobility which protects energy sources to organize. The aim of this paper is to evaluate various deployed strategies involving sink mobility. Multiple mobile sinks are capable of performing computational operations like collecting information from electric joints instantly, storage, and also communication capability. It evaluates the results and the effect of sink mobility by comparing with another routing protocol GEDAR.

ON THE APPLİCATİON OF THE BİNARY NUMBER SYSTEM İN CREATİNG QUESTİONS STRATEGY

It is obvious that the binary number system has found widespread application in modern computer technology in order to sufficiently simplify the computational logic operations. The simple principle of operation of the binary number system allows its application in areas other than computer technology. This article examines the application of the principles of binary numbers and information technology to simplify the questioning system. In this regard, the publication of this article is relevant.

GIS-SWIAS: Tool to Summarize Seawater Intrusion Status and Vulnerability at Aquifer Scale

In this paper, we introduce GIS-SWIAS, a novel generalized ArcGIS ArcToolbox that helps to analyze seawater intrusion (SWI) status and vulnerability at aquifer scale (SWIAS). It is a user-friendly tool that can be applied to any aquifer and is fully integrated in the ArcGIS environment, which is a widely available software tool. It is the first ArcGIS tool with these characteristics focusing on SWI analyses that we can find in the literature. GIS-SWIAS is able to deal with georeferenced information; it is easy to introduce the required data (inputs) and to efficiently perform the demanding computational operations required. Its outputs are in the form of shapes, reports, and images (maps, conceptual cross sections, and time series of lumped indices) to summarize the magnitude, intensity, and temporal evolution of SWI within an aquifer for specific dates or by showing statistics for a chosen time period. It can be applied to assess historical SWI dynamic in cases where there is no groundwater flow model. In those cases, the spatial distribution is assessed by applying simple interpolation techniques. Nevertheless, if we want a rational quantitative analysis of the sustainability of alternative management scenarios to the SWI problem, the GIS-SWIAS tool requires that information on hydraulic head and chloride concentration distribution is generated from simulations of their impacts by a calibrated density-dependent flow model. In such cases, adaptation strategies to potential future scenarios—whose distributed impacts have to be propagated within the previously calibrated models—could usefully be analyzed and compared using this tool. Given all these ways that the GIS-SWIAS tool can be applied, it provides a valuable tool for both the researcher and technician to assess SWI dynamics and aquifer resilience under different scenarios. It can support the decision-making process by helping to make a rational selection of sustainable management strategies. Its performance for the analyses of historical and potential future scenarios has been tested and confirmed in two case studies described in previous research works.

Toward Distributed, Global, Deep Learning Using IoT Devices

Deep learning (DL) using large scale, high-quality IoT datasets can be computationally expensive. Utilizing such datasets to produce a problem-solving model within a reasonable time frame requires a scalable distributed training platform/system. We present a novel approach where to train one DL model on the hardware of thousands of mid-sized IoT devices across the world, rather than the use of GPU cluster available within a data center. We analyze the scalability and model convergence of the subsequently generated model, identify three bottlenecks that are: high computational operations, time consuming dataset loading I/O, and the slow exchange of model gradients. To highlight research challenges for globally distributed DL training and classification, we consider a case study from the video data processing domain. A need for a two-step deep compression method, which increases the training speed and scalability of DL training processing, is also outlined. Our initial experimental validation shows that the proposed method is able to improve the tolerance of the distributed training process to varying internet bandwidth, latency, and Quality of Service metrics.

Совместное использование технологий OpenMP и MPI на узлах вычислительного кластера

This work is devoted to the problem of implementing an efficient parallel program that solves the asigned task using the maximum available amount of computing cluster resources in order to obtain the corresponding gain in performance with respect to the sequential version of the algorithm. The main objective of the work was to study the possibilities of joint use of the parallelization technologies OpenMP and MPI, considering the characteristics and features of the problems being solved, to increase the performance of executing parallel algorithms and programs on a computing cluster. This article provides a brief overview of approaches to calculating the sequential programs complexity functions. To determine the parallel programs complexity, an approach based on operational analysis was used. The features of the sequential programs parallelization technologies OpenMP and MPI are described. The main software and hardware factors affecting the execution speed of parallel programs on the nodes of a computing cluster are presented. The main attention in this paper is paid to the study of the impact on performance of computational and exchange operations number ratio in programs. To implement the research, parallel OpenMP and MPI testing programs were developed, in which the total number of operations and the correlation between computational and exchange operations are set. A computing cluster consisting of several nodes was used as a hardware and software platform. Experimental studies have made it possible to confirm the effectiveness of the hybrid model of a parallel program in multi-node systems with heterogeneous memory using OpenMP in shared memory subsystems, and MPI in a distributed memory subsystems

A computational-graph partitioning method for training memory-constrained DNNs

Many state-of-the-art Deep Neural Networks (DNNs) have substantial memory requirements. Limited device memory becomes a bottleneck when training those models. We propose ParDNN, an automatic, generic, and non-intrusive partitioning strategy for DNNs that are represented as computational graphs. ParDNN decides a placement of DNN’s underlying computational graph operations across multiple devices so that the devices’ memory constraints are met and the training time is minimized. ParDNN is completely independent of the deep learning aspects of a DNN. It requires no modification neither at the model nor at the systems level implementation of its operation kernels. ParDNN partitions DNNs having billions of parameters and hundreds of thousands of operations in seconds to few minutes. Our experiments with TensorFlow on 16 GPUs demonstrate efficient training of 5 very large models while achieving superlinear scaling for both the batch size and training throughput. ParDNN either outperforms or qualitatively improves upon the related work.

A Simple Algorithm for Compensation for Range Cell Migration in a Stripmap SAR

Introduction. Range Cell Migration (RCM) is a source of image blurring in synthetic aperture radars (SAR). There are two groups of signal processing algorithms used to compensate for migration effects. The first group includes algorithms that recalculate the SAR signal from the "along–track range – slant range" coordinate system into the "along-track range – cross-track range" coordinates using the method of interpolation. The disadvantage of these algorithms is their considerable computational cost. Algorithms of the second group do not rely on interpolation thus being more attractive in terms of practical application.Aim. To synthesize a simple algorithm for compensating for RCM without using interpolation.Materials and methods. The synthesis was performed using a simplified version of the Chirp Scaling algorithm.Results. A simple algorithm, which presents a modification of the Keystone Transform algorithm, was synthesized. The synthesized algorithm based on Fast Fourier Transforms and the Hadamard matrix products does not require interpolation.Conclusion. A verification of the algorithm quality via mathematical simulation confirmed its high efficiency. Implementation of the algorithm permits the number of computational operations to be reduced. The final radar image produced using the proposed algorithm is built in the true Cartesian coordinates. The algorithm can be applied for SAR imaging of moving targets. The conducted analysis showed that the algorithm yields the image of a moving target provided that the coherent processing interval is sufficiently large. The image lies along a line, which angle of inclination is proportional to the projection of the target relative velocity on the line-of-sight. Estimation of the image parameters permits the target movement parameters to be determined.

Advances in Atomtronics.

Atomtronics is a relatively new subfield of atomic physics that aims to realize the device behavior of electronic components in ultracold atom-optical systems. The fact that these systems are coherent makes them particularly interesting since, in addition to current, one can impart quantum states onto the current carriers themselves or perhaps perform quantum computational operations on them. After reviewing the fundamental ideas of this subfield, we report on the theoretical and experimental progress made towards developing externally-driven and closed loop devices. The functionality and potential applications for these atom analogs to electronic and spintronic systems is also discussed.

Efficient Certificate-Less Aggregate Signature Scheme with Conditional Privacy-Preservation for Vehicular Ad Hoc Networks Enhanced Smart Grid System.

Vehicular Ad hoc networks (VANETs) as spontaneous wireless communication technology of vehicles has a wide range of applications like road safety, navigation and other electric car technologies, however its practicability is greatly hampered by cyber-attacks. Due to message broadcasting in an open environment during communication, VANETs are inherently vulnerable to security and privacy attacks. However to address the cyber-security issues with optimal computation overhead is a matter of current security research challenge. So this paper designs a secure and efficient certificate-less aggregate scheme (ECLAS) for VANETs applicable in a smart grid scenario. The proposed scheme is based on elliptic curve cryptography to provide conditional privacy-preservation by incorporating usage of time validated pseudo-identification for communicating vehicles besides sorting out the KGC (Key Generation Center) escrow problem. The proposed scheme is comparatively more efficient to relevant related research work because it precludes expensive computation operations likes bilinear pairings as shown by the performance evaluation. Similarly, communication cost is within the ideal range to most related works while considering the security requirements of VANETs system applicable in a smart grid environment.

A scalable approach for the efficient segmentation of hyperspectral images

The number of applications of hyperspectral imaging (HSI) is steadily increasing, as technology evolves and cameras become more affordable. However, the volume of data in a hyperspectral image is large (order of Gigabytes) and standard off-the-shelf algorithms for multi-channel image analysis cannot be readily applied, due to the prohibitive computational time and large memory requirements. Therefore, new scalable approaches are required to perform hyperspectral image analysis. In this article we address an efficient methodology for conducting Unsupervised Image Segmentation – one of the basic and most fundamental image analysis operations. In the methodology proposed, unsupervised segmentation is conducted after transforming the spectral and spatial dimensions of the raw hyperspectral image into a more compact representation using multivariate and multiresolution techniques. The clusters identified in the compact image representation are then used to train a discriminative classifier. The classifier is then adapted and transferred for application to the raw image, where it will efficiently label all the original pixels. With the proposed methodology, the computational expensive operations (unsupervised clustering and classifier learning) are minimized, whereas the efficient implementation of the classifier guarantees the analysis at the native resolution. The effectiveness of the proposed methodology was tested on a real case study considering an industrial hyperspectral image capturing the reflectance spectrum for several objects made of different unknown materials. A significant reduction in the computational cost was achieved without compromising the quality of the unsupervised segmentation, demonstrating the potential of the proposed approach.

Large-scale neuromorphic optoelectronic computing with a reconfigurable diffractive processing unit

Application-specific optical processors have been considered disruptive technologies for modern computing that can fundamentally accelerate the development of artificial intelligence (AI) by offering substantially improved computing performance. Recent advancements in optical neural network architectures for neural information processing have been applied to perform various machine learning tasks. However, the existing architectures have limited complexity and performance; and each of them requires its own dedicated design that cannot be reconfigured to switch between different neural network models for different applications after deployment. Here, we propose an optoelectronic reconfigurable computing paradigm by constructing a diffractive processing unit (DPU) that can efficiently support different neural networks and achieve a high model complexity with millions of neurons. It allocates almost all of its computational operations optically and achieves extremely high speed of data modulation and large-scale network parameter updating by dynamically programming optical modulators and photodetectors. We demonstrated the reconfiguration of the DPU to implement various diffractive feedforward and recurrent neural networks and developed a novel adaptive training approach to circumvent the system imperfections. We applied the trained networks for high-speed classifying of handwritten digit images and human action videos over benchmark datasets, and the experimental results revealed a comparable classification accuracy to the electronic computing approaches. Furthermore, our prototype system built with off-the-shelf optoelectronic components surpasses the performance of state-of-the-art graphics processing units (GPUs) by several times on computing speed and more than an order of magnitude on system energy efficiency.

An efficient highly parallelized ReRAM-based architecture for motion estimation of HEVC

Motion estimation (ME) is a high efficiency video coding (HEVC) process for determining motion vectors that describe the blocks transformation direction from one adjacent frame to a future frame in a video sequence. ME is a memory and computation consuming process which accounts for more than 50% of the total running time of HEVC. To conquer the memory and computation challenges, this paper presents ReME, a highly paralleled processing-in-memory (PIM) architecture for the ME process based on resistive random access memory (ReRAM). In ReME, the space of ReRAM is mainly separated into storage engine and ME processing engine. The storage engine is used as conventional memory to store video frames and intermediate data, while the computation operations of ME are performed in ME processing engines. Each ME processing engine in ReME consists of Sum of Absolute Differences (SAD) modules, interpolation modules, and Sum of Absolute Transformed Difference (SATD) modules that transfer ME functions into ReRAM-based logic analog computation units. ReME further cooperates these basic computation units to perform ME processes in a highly parallel manner. Simulation results show that the proposed ReME accelerator significantly outperforms other implementations with time consuming and energy saving.

Trust Enforced Computational Offloading for Health Care Applications in Fog Computing

Internet of Things (IoT) is a network of internet connected devices that generates huge amount of data every day. The usage of IoT devices such as smart wearables, smart phones, smart cities are increasing in the linear scale. Health care is one of the primary applications today that uses IoT devices. Data generated in this application may need computation, storage and data analytics operations which requires resourceful environment for remote patient health monitoring. The data related with health care applications are primarily private and should be readily available to the users. Enforcing these two constraints in cloud environment is a hard task. Fog computing is an emergent architecture for providing computation, storage, control and network services within user’s proximity. To handle private data, the processing elements should be trustable entities in Fog environment. In this paper we propose novel Trust Enforced computation ofFLoading technique for trust worthy applications using fOg computiNg (TEFLON). The proposed system comprises of two algorithms namely optimal service offloader and trust assessment for addressing security and trust issues with reduced response time. And the simulation results show that proposed TEFLON framework improves success rate of fog collaboration with reduced average latency for delay sensitive applications and ensures trust for trustworthy applications.

Lightweight attention convolutional neural network through network slimming for robust facial expression recognition

Deep convolutional neural networks (DCNNs) have achieved outstanding results in facial expression recognition (FER). However, their runtime memory and computational resource requirements make it challenging to deploy them on resource-constrained devices, such as mobile devices. In this paper, we propose a novel lightweight attention DCNN (LA-Net) for robust FER, which uses squeeze-and-excitation (SE) modules and the network slimming strategy. First, we combine the SE modules with the CNN network, which assigns a certain weight to each feature channel. This enables LA-Net to focus on learning the prominent facial features, reduce redundant information, and finally extract discriminative features from facial images. Then, we use the network slimming method to further reduce the model’s size, which results in a thin and compact network that uses less runtime memory and computational operations with minimal accuracy loss. The proposed LA-Net model can achieve 95.52%, 87.00% and 100% test accuracy on KDEF, RAF-DB and FERG-DB FER datasets, respectively. The experimental results show that the proposed method achieves better or comparable results than state-of-the-art FER methods and significantly reduces the computational cost and the number of parameters, with better generalization capability and robustness.

Information Support Strategic Management Processes Environmental Policy of Cities

In the report presents some methodological and applied issues of information support for strategic management processes environmental policy of cities, which, within the framework of the National concept of sustainable development of Russia (1996) relate to the lowest level of its system organization. The purpose and the main direction of using the results is the development of “human-machine” planning systems (or support) for optimal management decisions for regional and local structures of environmental management. The research methodology based on the new paradigm of system representation of the world, the theory of “noosphere” by P. T. de Chardin, E. Le Roy (1927), V. I. Vernadsky (1933), a “new theory of entropy” by A. N. Panchenkov (1999). Object of research – the Active Complex organized Systems (ACS) class “nature-society”; subjects – concept and conditions of “equilibrium” and “Sustainable Development” (CSD in the abbreviation UN, 1993) such systems. Research methods – game-theoretic modeling and “Games against Nature”: antagonistic, coalition, cooperative, strategic, situational, etc., which required by the conditions of setting tasks, games. In the simulation, this will ensure the full integration of all currently available natural science knowledge, as well as the ability to perform computational operations others organizational management systems of the class “nature-society”. For that in this study suggests a new “geosystem” approach to the study of sustainable development problems in urban environmental policy, which includes the possibilities of both external (“technogenic”) and internal (or “soft”, that is, maintaining and restoring the ecological balance) management of Nature and Society.

Parallel algorithms for the creation of medical database

Analysis of parallel algorithms for graphics processors allows you to determine the bottlenecks of the algorithm that affect its performance on a particular computing system. Algorithms can be analyzed both at the AGM level and at the level of the CPU-GPU system. Despite the fact that the model does not reflect the overlap of the execution of computational operations and operations of access to the global memory, but estimates the upper limit of the execution time of the algorithm on the AGM, in any parallel algorithms for the AGM it is necessary to reduce this parameter of the algorithm due to the large number of clock cycles spent by the multiprocessor to implement global memory access operations.