Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “data visualization”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 973 records · Page 54

Damsel: A Data Model Storage Library for Exascale Science

Computational science applications have been described as having one of seven motifs (the “seven dwarfs”), each having a particular pattern of computation and communication. From a storage and I/O perspective, these applications can also be grouped into a number of data model motifs describing the way data is organized and accessed during simulation, analysis, and visualization. Major storage data models developed in the 1990s, such as Network Common Data Format (netCDF) and Hierarchical Data Format (HDF) projects, created support for more complex data models. Development of both netCDF and HDF5 was influenced by multi-dimensional dataset storage requirements, but their access models and formats were designed with sequential storage in mind (e.g., a POSIX I/O model). Although these and other high-level I/O libraries have had a beneficial impact on large parallel applications, they do not always attain a high percentage of peak I/O performance due to fundamental design limitations, and they do not address the full range of current and future computational science data models. The goal of this project is to enable exascale computational science applications to interact conveniently and efficiently with storage through abstractions that match their data models. The project consists of three major activities: (1) identifying major data model motifs in computational science applications and developing representative benchmarks; (2) developing a data model storage library, called Damsel, that supports these motifs, provides efficient storage data layouts, incorporates optimizations to enable exascale operation, and is tolerant to failures; and (3) productizing Damsel and working with computational scientists to encourage adoption of this library by the scientific community. The product of this project, Damsel library, is openly available for download from http://cucis.ece.northwestern.edu/projects/DAMSEL. Several case studies and application programming interface reference are also available to assist new users to learn to use the library.

97 MATHEMATICS AND COMPUTING↗

FY08 LDRD Final Report LOCAL: Locality-Optimizing Caching Algorithms and Layouts

This project investigated layout and compression techniques for large, unstructured simulation data to reduce bandwidth requirements and latency in simulation I/O and subsequent post-processing, e.g. data analysis and visualization. The main goal was to eliminate the data-transfer bottleneck - for example, from disk to memory and from central processing unit to graphics processing unit - through coherent data access and by trading underutilized compute power for effective bandwidth and storage. This was accomplished by (1) designing algorithms that both enforce and exploit compactness and locality in unstructured data, and (2) adapting offline computations to a novel stream processing framework that supports pipelining and low-latency sequential access to compressed data. This report summarizes the techniques developed and results achieved, and includes references to publications that elaborate on the technical details of these methods.

99 GENERAL AND MISCELLANEOUS↗

The development of a high-throughput gradient array apparatus for the study of porous polymer networks.

A gradient array apparatus was constructed for the study of porous polymers produced using the process of chemically-induced phase separation (CIPS). The apparatus consisted of a 60 element, two-dimensional array in which a temperature gradient was placed in the y-direction and composition was varied in the x-direction. The apparatus allowed for changes in opacity of blends to be monitored as a function of temperature and cure time by taking images of the array with time. The apparatus was validated by dispense a single blend composition into all 60 wells of the array and curing them for 24 hours and doing the experiment in triplicate. Variations in micron scale phase separation were readily observed as a function of both curing time and temperature and there was very good well-to-well consistency as well as trial-to-trial consistency. Poragen of samples varying with respect to cure temperature was removed and SEM images were obtained. The results obtained showed that cure temperature had a dramatic affect on sample morphology, and combining data obtained from visual observations made during the curing process with SEM data can enable a much better understanding of the CIPS process and provide predictive capability through the relatively facile generation of composition-process-morphology relationships. Data quality could be greatly enhanced by making further improvements in the apparatus. The primary improvements contemplated include the use of a more uniform light source, an optical table, and a CCD camera with data analysis software. These improvements would enable quantification of the amount of scattered light generated from individual elements as a function of cure time. In addition to the gradient array development, porous composites were produced by incorporating metal particles into a blend of poragen, epoxy resin, and crosslinker. The variables involved in the experiment were metal particle composition, primary metal particle size, metal concentration, and poragen composition. A total of 16 different porous composites were produced and characterized using SEM. In general, the results showed that pore morphology and the distribution of metal particles was dependent on multiple factors. For example, the use of silver nanoparticles did not significantly affect pore morphology for composites derived from decanol as the poragen, but exceptionally large pores were obtained with the use of decane as the poragen. With regard to the effect of metal particle size, silver nanoparticles were essentially exclusively dispered in the polymer matrix while silver microparticles were found in pores. For nickel particles, both nanoparticles and microparticles were largely dispersed in the polymer matrix and not in the pores.

36 MATERIALS SCIENCE↗

Variable Grid Method

NETL’s Variable Grid Method (VGM) is a novel approach that leverages GIS capabilities to simultaneously visualize and quantify spatial data trends and underlying data uncertainty. The flexible VGM approach utilizes a range of spatial datasets and uncertainty quantifications, which can be calculated using data related to sample density, sample variance, interpolation error, multiple simulations, etc., to create an integrated visualization of data and uncertainty. The intuitive manner of the VGM helps communicate the relationship between uncertainty and spatial data to effectively guide research, support advanced computation analyses, and helps inform research, management and policy decisions. The VGM approach is being developed into a standalone tool designed to integrate into a variety of research workflows, providing users the capabilities to utilize the VGM for their datasets, analysis, and models in support of their decision making needs. The VGM approach was developed as part of NETL’s Offshore portfolio. It was further matured and utilized in projects associated with both NETL’s Unconventional Resources and Carbon Storage portfolios, and is now being further developed by Varigrid, LLC.

Carbon Storage↗

Automated X-Ray Diffraction of Irradiated Materials

Synchrotron-based X-ray diffraction (XRD) and small-angle Xray scattering (SAXS) characterization techniques used on unirradiated and irradiated reactor pressure vessel steels yield large amounts of data. Machine learning techniques, including PCA, offer a novel method of analyzing and visualizing these large data sets in order to determine the effects of chemistry and irradiation conditions on the formation of radiation induced precipitates. In order to run analysis on these data sets, preprocessing must be carried out to convert the data to a usable format and mask the 2-D detector images to account for experimental variations. Once the data has been preprocessed, it can be organized and visualized using principal component analysis (PCA), multi-dimensional scaling, and k-means clustering. In conclusion, from these techniques, it is shown that sample chemistry has a notable effect on the formation of the radiation induced precipitates in reactor pressure vessel steels.

46 INSTRUMENTATION RELATED TO NUCLEAR SCIENCE AND ↗

Arrowland v1.0

"Arrowland" is a web-based software application primarily for mapping, integrating and visualizing a variety of metabolism data of living organisms, including but not limited to metabolomics, proteomics, transcriptomics and fluxomics. This software application makes multi-omics data analysis intuitive and interactive. It improves data sharing and communication by enabling users to visualize their omics data using a web browser (on a PC or mobile device). It increases user's productivity by simplifying multi-omics data analysis using well developed maps as a guide. Users using this tool can gain insights into their data sets that would be difficult or even impossible to tease out by looking at raw number, or using their currently existing toolchains to generate static single-use maps. Arrowland helps users save time by visualizing relative changes in different conditions or over time, and helps users to produce more significant insights faster. Preexisting maps decrease the learning curve for beginners in the omics field. Sets of multi-omics data are presented in the browser, as a two-dimensional flowchart resembling a map, with varying levels of detail information, based on the scaling of the map. Users can pan and zoom to explore different maps, compare maps, upload their own research data sets onto desired maps, alter map appearance in ways that facilitate interpretation, visualization and analysis of the given data, and export data, reports and actionable items to help the user initiative.

BIRKEL, GARRETT↗

A Structure-Based Distance Metric for High-Dimensional Space Exploration with Multi-Dimensional Scaling

Although the Euclidean distance does well in measuring data distances within high-dimensional clusters, it does poorly when it comes to gauging inter-cluster distances. This significantly impacts the quality of global, low-dimensional space embedding procedures such as the popular multi-dimensional scaling (MDS) where one can often observe non-intuitive layouts. We were inspired by the perceptual processes evoked in the method of parallel coordinates which enables users to visually aggregate the data by the patterns the polylines exhibit across the dimension axes. We call the path of such a polyline its structure and suggest a metric that captures this structure directly in high-dimensional space. This allows us to better gauge the distances of spatially distant data constellations and so achieve data aggregations in MDS plots that are more cognizant of existing high-dimensional structure similarities. Our MDS plots also exhibit similar visual relationships as the method of parallel coordinates which is often used alongside to visualize the high-dimensional data in raw form. We then cast our metric into a bi-scale framework which distinguishes far-distances from near-distances. The coarser scale uses the structural similarity metric to separate data aggregates obtained by prior classification or clustering, while the finer scale employs the appropriate Euclidean distance.

Clustering↗

Parallel In Situ Coupling of a Simulation with a Fully Featured Visualization System

There is a widening gap between compute performance and the ability to store computation results. Complex scientific codes are the most affected since they must save massive files containing meshes and fields for offline analysis. Time and storage costs instead dictate that data analysis and visualization be combined with the simulations themselves, being done in situ so data are transformed to a manageable size before they are stored. Earlier approaches to in situ processing involved combining specific visualization algorithms into the simulation code, limiting flexibility. We introduce a new library which instead allows a fully-featured visualization tool, VisIt, to request data as needed from the simulation and apply visualization algorithms in situ with minimal modification to the application code.

99 GENERAL AND MISCELLANEOUS↗

GIS Visualization of Transportation Energy Consumption

Transportation is witnessing unprecedented transformation, where emerging technologies are disrupting the way we travel. Be it sharing economy, or micro-mobility, the landscape of urban transportation is undergoing a much-needed paradigm shift. In order to capture and model these shifts, researchers need to be agile in their studies of the current and future transportation landscape. The need for agility in turn calls for sophisticated, intuitive, and reliable means to visualize and share transportation data. Travel is inherently spatial, so methods of visualizing transportation information should also be rooted in spatial analysis. Geographic Information Systems (GIS) are the premier technological framework to analyze and display spatial data. GIS tools have been used in the past to depict flow of vehicles, and display of network conditions (speed, congestion, etc.). However, in an ever-changing technological landscape, spanning advancements in vehicle as well as information systems, depicting vehicle flows falls short of providing a comprehensive picture of the impact these advancements have on travel related energy consumption. To address this issue, this research effort presents a web-based mapping application to visualize how travel related energy flows across a city. The application is being developed by National Renewable Energy Laboratory researchers to integrate various transportation energy consumption models within a GIS schema. This has the goal of enabling rapid analysis of energy impacts of the dramatically evolving transportation environment. The application is being developed using a Python framework for ArcGIS Online. Using road network data from the City of Columbus, Ohio and traffic data from CATT Laboratory's Regional Integrated Transportation Information System, transportation energy consumption will be modeled at a macro level. The application aims to provide highly accurate data while still maintaining the flexibility needed to adapt the model when new technologies arise. The methodology presented through this effort is expected to provide insights into how light-duty vehicles use energy on a large scale based on the road network they use. The application will also provide data exports to enable sharing and collaboration. " need for agility in turn calls for sophisticated, intuitive, and reliable means to visualize and share transportation data. Travel is inherently spatial, so methods of visualizing transportation information should also be rooted in spatial analysis. Geographic Information Systems (GIS) are the premier technological framework to analyze and display spatial data. GIS tools have been used in the past to depict flow of vehicles, and display of network conditions (speed, congestion, etc.). However, in an ever-changing technological landscape, spanning advancements in vehicle as well as information systems, depicting vehicle flows falls short of providing a comprehensive picture of the impact these advancements have on travel related energy consumption. To address this issue, this research effort presents a web-based mapping application to visualize how travel related energy flows across a city. The application is being developed by National Renewable Energy Laboratory researchers to integrate various transportation energy consumption models within a GIS schema. This has the goal of enabling rapid analysis of energy impacts of the dramatically evolving transportation environment. The application is being developed using a Python framework for ArcGIS Online. Using road network data from the City of Columbus, Ohio and traffic data from CATT Laboratory's Regional Integrated Transportation Information System, transportation energy consumption will be modeled at a macro level. The application aims to provide highly accurate data while still maintaining the flexibility needed to adapt the model when new technologies arise. The methodology presented through this effort is expected to provide insights into how light-duty vehicles use energy on a large scale based on the road network they use. The application will also provide data exports to enable sharing and collaboration.

33 ADVANCED PROPULSION SYSTEMS↗

Optimizing automatic morphological classification of galaxies with machine learning and deep learning using Dark Energy Survey imaging

There are several supervised machine learning methods used for the application of automated morphological classification of galaxies; however, there has not yet been a clear comparison of these different methods using imaging data, or an investigation for maximizing their effectiveness. We carry out a comparison between several common machine learning methods for galaxy classification [Convolutional Neural Network (CNN), K-nearest neighbour, logistic regression, Support Vector Machine, Random Forest, and Neural Networks] by using Dark Energy Survey (DES) data combined with visual classifications from the Galaxy Zoo 1 project (GZ1). Our goal is to determine the optimal machine learning methods when using imaging data for galaxy classification. We show that CNN is the most successful method of these ten methods in our study. Using a sample of ~2800 galaxies with visual classification from GZ1, we reach an accuracy of ~0.99 for the morphological classification of ellipticals and spirals. The further investigation of the galaxies that have a different ML and visual classification but with high predicted probabilities in our CNN usually reveals the incorrect classification provided by GZ1. We further find the galaxies having a low probability of being either spirals or ellipticals are visually lenticulars (S0), demonstrating that supervised learning is able to rediscover that this class of galaxy is distinct from both ellipticals and spirals. We confirm that ~2.5 percent galaxies are misclassified by GZ1 in our study. After correcting these galaxies’ labels, we improve our CNN performance to an average accuracy of over 0.99 (accuracy of 0.994 is our best result).

79 ASTRONOMY AND ASTROPHYSICS↗

CMed: Crowd Analytics for Medical Imaging Data

In this work, we present a visual analytics framework, CMed, for exploring medical image data annotations acquired from crowdsourcing. CMed can be used to visualize, classify, and filter crowdsourced clinical data based on a number of different metrics such as detection rate, logged events, and clustering of the annotations. CMed provides several interactive linked visualization components to analyze the crowd annotation results for a particular video and the associated workers. Additionally, all results of an individual worker can be inspected using multiple linked views in our CMed framework. We allow a crowdsourcing application analyst to observe patterns and gather insights into the crowdsourced medical data, helping him/her design future crowdsourcing applications for optimal output from the workers. We demonstrate the efficacy of our framework with two medical crowdsourcing studies: polyp detection in virtual colonoscopy videos and lung nodule detection in CT thin-slab maximum intensity projection videos. We also provide experts' feedback to show the effectiveness of our framework. Lastly, we share the lessons we learned from our framework with suggestions for integrating our framework into a clinical workflow.

97 MATHEMATICS AND COMPUTING↗

A Practical Solver for Scalar Data Topological Simplification

This paper presents a practical approach for the optimization of topological simplification, a central pre-processing step for the analysis and visualization of scalar data. Given an input scalar field f and a set of “signal” persistence pairs to maintain, our approaches produces an output field g that is close to f and which optimizes (i) the cancellation of “non-signal” pairs, while (ii) preserving the “signal” pairs. In contrast to pre-existing simplification algorithms, our approach is not restricted to persistence pairs involving extrema and can thus address a larger class of topological features, in particular saddle pairs in three-dimensional scalar data. Our approach leverages recent generic persistence optimization frameworks and extends them with tailored accelerations specific to the problem of topological simplification. Extensive experiments report substantial accelerations over these frameworks, thereby making topological simplification optimization practical for real-life datasets. Our approach enables a direct visualization and analysis of the topologically simplified data, e.g., via isosurfaces of simplified topology (fewer components and handles). We apply our approach to the extraction of prominent filament structures in three-dimensional data. Specifically, we show that our pre-simplification of the data leads to practical improvements over standard topological techniques for removing filament loops. Here, we also show how our approach can be used to repair genus defects in surface processing. Finally, we provide a C++ implementation for reproducibility purposes.

97 MATHEMATICS AND COMPUTING↗

A combined static-dynamic single-dose imaging protocol to compare quantitative dynamic SPECT with static conventional SPECT

Background: SPECT myocardial perfusion imaging (MPI) is a clinical mainstay that is typically performed with static imaging protocols and visually or semi-quantitatively assessed for perfusion defects based upon the relative intensity of myocardial regions. Dynamic cardiac SPECT presents a new imaging technique based on time-varying information of radiotracer distribution, which permits the evaluation of regional myocardial blood flow (MBF) and coronary flow reserve (CFR). In this work, a preliminary feasibility study was conducted in a small patient sample designed to implement a unique combined static-dynamic single-dose one-day visit imaging protocol to compare quantitative dynamic SPECT with static conventional SPECT for improving the diagnosis of coronary artery disease (CAD). Methods: Fifteen patients (11 males, four females, mean age 71 ± 9 years) were enrolled for a combined dynamic and static SPECT (Infinia Hawkeye 4, GE Healthcare) imaging protocol with a single dose of {sup 99m}Tc-tetrofosmin administered at rest and a single dose administered at stress in a one-day visit. Out of 15 patients, eleven had selective coronary angiography (SCA), 8 within 6 months and the rest within 24 months of SPECT imaging, without intervening symptoms or interventions. The extent and severity of perfusion defects in each myocardial region was graded visually. Dynamically acquired data were also used to estimate the MBF and CFR. Both visually graded images and estimated CFR were tested against SCA as a reference to evaluate the validity of the methods. Results: Overall, conventional static SPECT was normal in ten patients and abnormal in five patients, dynamic SPECT was normal in 12 patients and abnormal in three patients, and CFR from dynamic SPECT was normal in nine patients and abnormal in six patients. Among those 11 patients with SCA, conventional SPECT was normal in 5, 3 with documented CAD on SCA with an overall accuracy of 64%, sensitivity of 40% and specificity of 83%. Dynamic SPECT image analysis also produced a similar accuracy, sensitivity, and specificity. CFR was normal in 6, each with CAD on SCA with an overall accuracy of 91%, sensitivity of 80%, and specificity of 100%. The mean CFR was significantly lower for SCA detected abnormal than for normal patients (3.86±1.06 vs 1.94±0. 0.67, P < 0.001). Conclusions: The visually assessed image findings in static and dynamic SPECT are subjective, and may not reflect direct physiologic measures of coronary lesion based on SCA. The CFR measured with dynamic SPECT is fully objective, with better sensitivity and specificity, available only with the data generated from the dynamic SPECT method.

62 RADIOLOGY AND NUCLEAR MEDICINE↗

GPGPU Enabled Adaptive Volume Visualization Using Commodity Game Engines

The amount of data produced by researchers in fields like nuclear engineering, materials science, and medicine has dramatically increased in recent years. There is a growing need to analyze and visualize this data in order to derive new insights. However, creating a visualization of data that is many gigabytes in size in real time is a difficult task since most graphics processing units (GPUs) are unable to hold this much data at one time. In order to alleviate this issue, an adaptive volume visualizer has been developed that is capable of rendering large scale datasets with high levels of detail at interactive framerates.

97 MATHEMATICS AND COMPUTING↗

Interactive Visualization of Multi-Group Cross Sections on High-Fidelity Spatial Meshes

Core simulation tools utilizing Monte Carlo transport methods represent the 'Holy Grail' for nuclear reactor analysis. However, deterministic transport methods pose a number of advantages to Monte Carlo from a computational perspective, which seems likely to make these methods more viable in the near to intermediate term. Deterministic methods almost universally discretize the energy domain in a few to hundreds of energy groups. One of the key challenges for deterministic transport-based methods to be viable is accurate, reactor agnostic multi-group cross section generation. This paper describes tools for data processing and visualization of multigroup cross sections generated on high-fidelity spatial meshes from the OpenMC Monte Carlo code. Next generation deterministic transport reactor core modeling tools will require multi-group cross section (MGXS) generation on high-fidelity spatial meshes. High-fidelity MGXS are necessary for the accurate simulation of neutron physics in nuclear reactors with a high degree of spectral heterogeneity - e.g., large localized neutron flux gradients throughout a reactor core. Although many local (e.g., pin cell) approximations have been developed to model the scalar flux to compute spectral-averaged MGXS, these techniques cannot account for complex intra-pin and intra-assembly spatial self-shielding effects. Monte Carlo presents the most accurate method for reactor agnostic multi-group cross section generation since it does not require the use of local approximations to the flux. Monte Carlo has been proposed as a highly desirable framework for reactor agnostic cross section generation for many years. Monte Carlo methods can effectively embed all of the relevant spectral effects into the MGXS through the random particle tracking and interaction sampling processes. Previous work has applied MCNP and Serpent to compute few-group constants for diffusion calculations, and OpenMC to compute multi-group cross sections for transport-based methods. The primary difficulty with using Monte Carlo for MGXS generation is the tremendous computational resources required to adequately converge multi-group cross sections in each spatial region and energy group. Recent development of the OpenMC Monte Carlo code for high-performance computing platforms has made multigroup cross section generation from Monte Carlo more tractable in the foreseeable future. The recent release of OpenMC v0.7.1 includes a new Python-based module for MGXS generation. The MGXS generation module leverages many of the algorithms and features for 'big data' analysis of large tally datasets on complex spatial meshes. This paper describes the data management workflow and visualization tools which have been developed to explore the plethora of MGXS data generated with OpenMC. It is the authors' hope that the development of the visualization tool described in this paper will ultimately enable the design of machine learning algorithms to accelerate the convergence rate of high-fidelity multi-group cross section generated with Monte Carlo. (authors)

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

Beyond Visual Analytics: Human-Machine Teaming for AI-Driven Data Sensemaking

"Detect the expected, discover the unexpected" was the founding principle of the field of visual analytics. This mantra implies that human stakeholders, like a domain expert or data analyst, could leverage visual analytics techniques to seek answers to known unknowns and discover unknown unknowns in the course of the data sensemaking process. We argue that in the era of AI-driven automation, we need to recalibrate the roles of humans and machines (e.g., a machine learning model) as teammates. We posit that by realizing human-machine teams as a stakeholder unit, we can better achieve the best of both worlds: automation transparency and human reasoning efficacy. However, this also increases the burden on analysts and domain experts towards performing more cognitively demanding tasks than what they are used to. In this paper, we reflect on the complementary roles in a human-machine team through the lens of cognitive psychology and map them to existing and emerging research in the visual analytics community. We discuss open questions and challenges around the nature of human agency and analyze the shared responsibilities in human-machine teams.

Sensemaking, human-machine teaming, agency, artifi↗

Self-Organizing Maps and Their Applications to Data Analysis

Self-Organizing Maps(SOMs) are a form of unsupervised neural network that are used for visualization and exploratory data analysis of high dimensional datasets. Our goal was to understand how we can use a SOM to gain insights about datasets. We do this by first understanding the initialization, training, error metrics, and convergence properties of the SOM. Next we discuss the ways to interpret and visualize a Self-Organizing Map. Finally we used real datasets to understand what the Self-Organizing Map can tell us about labeled and unlabeled data. Based on experiments with our datasets we found that the Self-Organizing Map can tell us about the spacing and position of high dimensional clusters, help us find non-linear patterns, and give us insight into the shape of our data.

97 MATHEMATICS AND COMPUTING↗

Adaptive elasticity policies for staging-based in situ visualization

In situ processing aims to alleviate the growing gap between computation and I/O capabilities by performing data processing close to the data source. In situ processing is widely used to process data generated by multiple data sources, including observation data from edge devices or scientific observational facilities and the simulation data generated by scientific computation on a high-performance computing (HPC) platform. For a scientific workflow that is run on an HPC platform and composed of a simulation program and an in situ data analytics or visualization (abbreviated as ana/vis) task, there is an implicit assumption that the computing resources assigned to the workflow keep static during the workflow execution. However, with the converging trend between the HPC and cloud computing platform, running the in situ ana/vis task in an elastic way is promising to decrease its overhead and improve its resource utilization rate. Resource elasticity represents the ability to change resource configurations such as the number of computing nodes/processes during workflow execution. An elastic job may dynamically adjust resource configurations; it may use a few resources at the beginning and more resources toward the end of the job when interesting data appear. However, it is hard to predict a priori how many computing nodes/processes need to be added/removed during the workflow execution to adapt to changing workflow needs. How to efficiently guide elasticity operations, such as growing or shrinking the number of processes used for in situ analysis during workflow execution, is an open-ended research question. In this article, we present adaptive elasticity policies that adopt workflow runtime information collected during workflow execution to predict how to trigger the addition/removal of processes in order to minimize in situ processing overhead. Taking in situ visualization tasks as an example, we integrate the presented elasticity policies into a staging-based elastic workflow and evaluate its efficiency in multiple elasticity scenarios. Compared with the situation without elasticity or with a static elasticity policy that uses a fixed number of processes for each rescaling operation, the adaptive elasticity policy can save overhead in finding a proper resource configuration and improve resource utilization efficiency. Furthermore, one experiment illustrates that the adaptive elasticity policy saves 41% of core-hours compared with the situation without the resource elasticity.

97 MATHEMATICS AND COMPUTING↗