Adaptive Performance Model Research Articles

The Adaptive Performance Model of Marines: The Role of Job Demands and Crafting Strategy

The success of the Marine Corps organization depends on the ability of its personnel to respond positively and overcome environmental challenges, acquire the new required skills, take official orders assiduously, and exhibit the adaptive performance from assigned tasks. Therefore, this study aims to analyze the adaptive performance model suitable for Marine Corps personnel by using job demands and crafting strategy as predictor and mediator variables, respectively. A quantitative method was used with cross-sectional and purposive sampling techniques. The model was tested using a sample of 230 Marine Intermediate Officers. The goodness of fit of the adaptive performance model is quite positive, with RMR = .05, CFI = .926, and TLI= .908. The results of the Structural Equation Model (SEM) analysis with AMOS 22 showed that: 1) the adaptive performance model offered for Marine Intermediate Officers is quite fit. 2) Job demands have a strong direct effect on adaptive performance. 3) Job crafting acts as a partial mediator between work demands and adaptive performance.

Developing adaptive performance: A conceptual model to guide simulation-based training design.

Effective emergency department care requires individuals and teams to adapt to changes in patient condition, team factors, environmental issues, and system-level challenges. Adaptability is often listed as an important skill for emergency medicine physicians; however, conceptual models describing the processes involved in adaptive performance have not been translated for health care settings. Similarly, educators have not described training design strategies that support the development of adaptive performance. We examined the team science and health care literatures for key concepts in adaptive performance, health care team performance, and diagnostic decision-making. Using expert consensus, we integrated these concepts to develop the team adaptive performance model and to identify training design approaches that support the development of adaptability. We identify nine training principles supported by the team adaptive performance model and the adaptive learning system. Each training principle is accompanied by recommendations and mechanisms for implementation in emergency medicine simulation-based education. Training experiences can be designed to target processes that support adaptive performance.

Assessment of Dynamic Bayesian Models for Gas Turbine Diagnostics, Part 2: Discrimination of Gradual Degradation and Rapid Faults

There are many challenges that an effective diagnostic system must overcome for successful fault diagnosis in gas turbines. Among others, it has to be robust to engine-to-engine variations in the fleet, it has to discriminate between gradual deterioration and abrupt faults, and it has to identify sensor faults correctly and be robust in case of such faults. To combine their benefits and overcome their limitations, two diagnostic methods were integrated in this work to form a multi-layer system. An adaptive performance model was used to track gradual deterioration and detect rapid or abrupt anomalies, while a series of static and dynamic Bayesian networks were integrated to identify component degradation, component abrupt faults, and sensor faults. The proposed approach was tested on synthetic data and field data from a single-shaft gas turbine of 50 MW class. The results showed that the approach could give acceptable accuracy in the isolation and identification of multiple faults, with 99% detection and isolation accuracy and 1% maximum error in the identified fault magnitude. The approach was also proven robust to sensor faults, by replacing the faulty signal with an estimated value that had only 3% error compared to the real measurement.

Adaptive Performance Modeling of Data-intensive Workloads for Resource Provisioning in Virtualized Environment

The processing of data-intensive workloads is a challenging and time-consuming task that often requires massive infrastructure to ensure fast data analysis. The cloud platform is the most popular and powerful scale-out infrastructure to perform big data analytics and eliminate the need to maintain expensive and high-end computing resources at the user side. The performance and the cost of such infrastructure depend on the overall server configuration, such as processor, memory, network, and storage configurations. In addition to the cost of owning or maintaining the hardware, the heterogeneity in the server configuration further expands the selection space, leading to non-convergence. The challenge is further exacerbated by the dependency of the application’s performance on the underlying hardware. Despite an increasing interest in resource provisioning, few works have been done to develop accurate and practical models to proactively predict the performance of data-intensive applications corresponding to the server configuration and provision a cost-optimal configuration online. In this work, through a comprehensive real-system empirical analysis of performance, we address these challenges by introducing ProMLB: a proactive machine-learning-based methodology for resource provisioning. We first characterize diverse types of data-intensive workloads across different types of server architectures. The characterization aids in accurately capture applications’ behavior and train a model for prediction of their performance. Then, ProMLB builds a set of cross-platform performance models for each application. Based on the developed predictive model, ProMLB uses an optimization technique to distinguish close-to-optimal configuration to minimize the product of execution time and cost. Compared to the oracle scheduler, ProMLB achieves 91% accuracy in terms of application-resource matching. On average, ProMLB improves the performance and resource utilization by 42.6% and 41.1%, respectively, compared to baseline scheduler. Moreover, ProMLB improves the performance per cost by 2.5× on average.

Probabilistic Model for Aero-Engines Fleet Condition Monitoring

Since aeronautic transportation is responsible for a rising share of polluting emissions, it is of primary importance to minimize the fuel consumption any time during operations. From this perspective, continuous monitoring of engine performance is essential to implement proper corrective actions and avoid excessive fuel consumption due to engine deterioration. This requires, however, automated systems for diagnostics and decision support, which should be able to handle large amounts of data and ensure reliability in all the multiple conditions the engines of a fleet can be found in. In particular, the proposed solution should be robust to engine-to-engine deviations and different sensors availability scenarios. In this paper, a probabilistic Bayesian network for fault detection and identification is applied to a fleet of engines, simulated by an adaptive performance model. The combination of the performance model and the Bayesian network is also studied and compared to the probabilistic model only. The benefit in the suggested hybrid approach is identified as up to 50% higher accuracy. Sensors unavailability due to manufacturing constraints or sensor faults reduce the accuracy of the physics-based method, whereas the Bayesian model is less affected.

An adaptive breadth-first search algorithm on integrated architectures

In the big data era, graph applications are becoming increasingly important for data analysis. Breadth-first search (BFS) is one of the most representative algorithms; therefore, accelerating BFS using graphics processing units (GPUs) is a hot research topic. However, due to their random data access pattern, it is difficult to take full advantage of the power of GPUs. Recently, hardware designers have integrated CPUs and GPUs on the same chip, allowing both devices to share physical memory, which provides the convenience of switching between CPUs and GPUs with little cost. BFS processing can be divided into several levels, and various traversal orders can be used at each level. Using different traversal orders on different devices (CPUs or GPUs) results in diverse performances. Thus, the challenge in using BFS on integrated architectures is how to select the traversal order and the device for each level. Previous works have failed to address this problem effectively. In this study, we propose an adaptive performance model that automatically finds a suitable traversal order and device for each level. We evaluated our method on Graph500, where it not only shows the best energy efficiency but also achieves a giga-traversed edges per second (GTEPS) performance of approximately 2.1 GTEPS, which is a $$2.3\,\times $$ speed improvement over the state-of-the-art BFS on integrated architectures.

Adaptive performance model for dynamic scaling Apache Spark Streaming

Nowadays, data volumes increase exceptionally, a lot of information comes from different sources, for example, from mobile phones, sensors, traffic, etc. All information from these sources can be represented as a data streams, which can grow up and fall in time in their size. In the first case, data processing requires optimization via dynamic resource allocation in order to decrease processing time, in the second case, it requires optimization related with resources deallocation because removing unnecessary resources can decrease the total cost. The question is how to identify optimal amount of resources to satisfy required processing delay under certain volume of workload? Current implementation of Apache Spark Streaming and existing models can’t give us such possibility. In this paper, we propose adaptive performance model, which can dynamically scale up and down Apache Spark Streaming platform on the AWS.

Progressive performance modeling for the strategic determinants of market value in the high-tech oriented SMEs

The purpose of this paper is to present an adaptive performance model using neural networks in scrutinizing the impact of strategic factors on firm performance, especially within high-tech oriented small and medium-sized enterprises (SMEs) in the United States. This paper explores generalized learning of backpropagation neural network (BPNN) in conducting an explanatory and predictive analysis of the strategic determinants of the market value of SMEs. The progressive performance model through BPNN is designed to capture the different and unique significance of strategic determinants for better firm performance by dividing high-tech segments into two performance groups: high performers and low performers. In doing so, this paper introduces a salient BPNN approach for performance modeling and extends the applications of BPNN. Furthermore, efficiency measurement and performance prediction using BPNN adds meaningful value to the literature and highlights the potential advantages of using BPNN. The empirical results demonstrate the successful implementation of the model and clearly distinguish varying patterns at different performance levels, High and Low, which is a significant finding of this study. Overall, sales growth, R&D intensity, and current ratio can be used as major strategic determinants of market value performance of the technology-oriented SMEs.

An integrated nonlinear model-based approach to gas turbine engine sensor fault diagnostics

Aircraft engine sensor fault diagnosis is closely related technology that assists operators in managing the health of gas turbine engine assets. As all gas turbine engines will exhibit performance changes due to usage, the on-board engine model built up initially will no longer track the engine over the course of the engine’s life, and then the model-based method for sensor fault diagnosis tends to be failure. This necessitates the study of the sensor fault diagnosis techniques due to usage over its operating life. Based on our recent results, an integrated approach based on nonlinear on-board model is developed for the gas turbine engine sensor fault diagnostics in this paper. The architecture is mainly composed of dual nonlinear engine models; one is a nonlinear real-time adaptive performance model and the other a nonlinear on-board baseline model. The extended Kalman filter estimator in the nonlinear real-time adaptive performance model is used to obtain the real-time estimates of component performance, and the nonlinear on-board baseline model with performance periodically update to provide the nominal reference in flight. The novel update strategy to sensor fault threshold based on the model errors and noise level is also presented. Important results are obtained on step fault and pulse fault behavior of the engine sensor. The proposed approach is easy to design and tune with long-term engine health degradation. Finally, experiment studies are provided to validate the benefit of the engine sensor fault diagnostics.

An adaptive performance modeling tool for GPU architectures

This paper presents an analytical model to predict the performance of general-purpose applications on a GPU architecture. The model is designed to provide performance information to an auto-tuning compiler and assist it in narrowing down the search to the more promising implementations. It can also be incorporated into a tool to help programmers better assess the performance bottlenecks in their code. We analyze each GPU kernel and identify how the kernel exercises major GPU microarchitecture features. To identify the performance bottlenecks accurately, we introduce an abstract interpretation of a GPU kernel, work flow graph , based on which we estimate the execution time of a GPU kernel. We validated our performance model on the NVIDIA GPUs using CUDA (Compute Unified Device Architecture). For this purpose, we used data parallel benchmarks that stress different GPU microarchitecture events such as uncoalesced memory accesses, scratch-pad memory bank conflicts, and control flow divergence, which must be accurately modeled but represent challenges to the analytical performance models. The proposed model captures full system complexity and shows high accuracy in predicting the performance trends of different optimized kernel implementations. We also describe our approach to extracting the performance model automatically from a kernel code.

Adaptive optics performance model for optical interferometry

The optical interferometry community has discussed the possibility of using adaptive optics (AO) on apertures much larger than the atmospheric coherence length in order to increase the sensitivity of an interferometer, although few quantitative models have been investigated. The aim of this paper is to develop an analytic model of an AO-equipped interferometer and to use it to quantify, in relative terms, the gains that may be achieved over an interferometer equipped only with tip-tilt correction. Functional forms are derived for wavefront errors as a function of spatial and temporal coherence scales and flux and applied to the AO and tip-tilt cases. In both cases, the AO and fringe detection systems operate in the same spectral region, with the sharing ratio and subaperture size as adjustable parameters, and with the interferometer beams assumed to be spatially filtered after wavefront correction. It is concluded that the use of AO improves the performance of the interferometer in three ways. First, at the optimal aperture size for a tip-tilt system, the AO system is as much as ~50% more sensitive. Second, the sensitivity of the AO system continues to improve with increasing aperture size. And third, the signal-to-noise ratio of low-visibility fringes in the bright-star limit is significantly improved over the tip-tilt case.

Identifying Faults in the Variable Geometry System of a Gas Turbine Compressor

The influence of faults in the variable geometry (variable stator vanes) system of a multistage axial compressor on the performance of an industrial gas turbine is investigated. An experimental investigation has been conducted, by implanting such faults into an operating gas turbine. The faults examined are individual stator vane mistunings of different magnitudes and located at different stages. Fault identification is based on the aerothermodynamic measurement data and is achieved by employing two different techniques, namely adaptive performance modeling and monitoring the circumferential distribution of the turbine exit temperature. It is observed that the deviations of the health indices produced by an adaptive performance model form patterns that can be used to identify the faults. The patterns characterize both the kind and the magnitude of the fault. On the other hand, the turbine exit temperature profile is also influenced and its change can be used as additional information, to increase the confidence level of the diagnosis (contrary to customary practice, which expects temperatures profiles to reflect only burner or turbine malfunctions).

Optimal Measurement and Health Index Selection for Gas Turbine Performance Status and Fault Diagnosis

In this paper, we present a method for defining the health estimation parameters and the measurements that must be used when a monitoring system for an engine is being set up. The particular engine layout, the available measuring instruments, and the accuracy by which data can be collected are the factors taken into account. The particular health condition estimation factors that have to be used are defined as a function of this information and the desired depth of fault identification. A fast selection procedure based on the method of singular value decomposition is presented. The uncertainty in the estimations is also derived, thus giving an additional element of information useful for decision making. The proposed method, together with adaptive performance modeling, provides a self-sufficient tool, which can be applied for setting up and subsequent exploitation of a health monitoring expert system. The advantage of the procedure is that it provides a frame of application, allowing quick implementation in a new engine of interest, other than the ones previously considered.