Development of an operational fitness test for the Royal Air Force

Since 2002, the Royal Air Force (RAF) has been working towards developing role-related physical tests for use as an operational fitness test (OFT). The purpose of this study was to establish reliability of the OFT (comprising four tests), investigate gym-based tests as predictors of performance and establish performance standards. Fifty-eight RAF personnel performed the OFT on three occasions. A separate cohort carried out fitness and anthropometric tests before performing the OFT, by way of establishing performance predictors. Documented evidence and views of an expert panel were used to determine OFT standards. Reliability ranged from moderate to good for three tests, with one test (Dig) showing poor reliability. The 95% limits of agreement for the prediction models ranged from good to poor (6.7-34.2%). The prediction models were not sufficiently accurate to estimate confidently OFT performance, but could be used as a guide to quantify likely outcome and training needs.     

Development of physical selection procedures for the British Army. Phase 2: Relationship between physical performance tests and criterion tasks

This paper is the second in a series of three to describe the development of physical selection standards for the British Army. The first paper defined criterion tasks (single lift, carry, repetitive lift and carry, and loaded march tasks) and set standards on the criterion tasks for all British Army trades. The principal objective was to determine which combination of physical performance tests could be best used to predict criterion task performance. Secondary objectives included developing so-called ‘gender-free’ and ‘gender-unbiased’ models. The objectives were met by analysing performance data on the criterion tasks and a large battery of physical performance tests collected from 379 trained soldiers (mean age 23.5 (SD 4.45) years, stature 1734 (SD 79.5) mm, body mass 71.4 (SD 10.58) kg). Objective 1 was met: the most predictive physical performance tests were identified for all criterion tasks. Both single lift tasks were successfully modelled using muscle strength and fat free mass scores. The carry model incorporated muscle endurance and body size data, but the errors of prediction were large. The repetitive lift models included measures of muscle strength and endurance, and body size, but errors of prediction were also large. The loaded march tasks were successfully modelled incorporating indices of aerobic fitness, supplemented by measures of strength, endurance or body size and composition. The secondary objectives were partially fulfilled, though limitations in the data hampered the process. Although only one model (a loaded march) was gender-free, three models were gender-related (i.e. contained ‘gender’ explicitly in the model). The remaining six were gender-specific (i.e. were appropriate for men or for women). Owing to both a lower accuracy of prediction in women's scores and a greater tendency for the women's scores to be distributed around the pass standards, a greater percentage of women than men were misclassified as passing or failing, resulting in indirect discrimination. A validation of the models in a separate sample of the user population of recruits is reported in the third paper in this series.     

Development of physical selection procedures for the British Army. Phase 2: relationship between physical performance tests and criterion tasks

This paper is the second in a series of three to describe the development of physical selection standards for the British Army. The first paper defined criterion tasks (single lift, carry, repetitive lift and carry, and loaded march tasks) and set standards on the criterion tasks for all British Army trades. The principal objective was to determine which combination of physical performance tests could be best used to predict criterion task performance. Secondary objectives included developing so-called 'gender-free' and 'gender-unbiased' models. The objectives were met by analysing performance data on the criterion tasks and a large battery of physical performance tests collected from 379 trained soldiers (mean age 23.5 (SD 4.45) years, stature 1734 (SD 79.5) mm, body mass 71.4 (SD 10.58) kg). Objective 1 was met: the most predictive physical performance tests were identified for all criterion tasks. Both single lift tasks were successfully modelled using muscle strength and fat free mass scores. The carry model incorporated muscle endurance and body size data, but the errors of prediction were large. The repetitive lift models included measures of muscle strength and endurance, and body size, but errors of prediction were also large. The loaded march tasks were successfully modelled incorporating indices of aerobic fitness, supplemented by measures of strength, endurance or body size and composition. The secondary objectives were partially fulfilled, though limitations in the data hampered the process. Although only one model (a loaded march) was gender-free, three models were gender-related (i.e. contained 'gender' explicitly in the model). The remaining six were gender-specific (i.e. were appropriate for men or for women). Owing to both a lower accuracy of prediction in women's scores and a greater tendency for the women's scores to be distributed around the pass standards, a greater percentage of women than men were misclassified as passing or failing, resulting in indirect discrimination. A validation of the models in a separate sample of the user population of recruits is reported in the third paper in this series.     

Coevolution of Fitness Predictors

We present an algorithm that coevolves fitness predictors, optimized for the solution population, which reduce fitness evaluation cost and frequency, while maintaining evolutionary progress. Fitness predictors differ from fitness models in that they may or may not represent the objective fitness, opening opportunities to adapt selection pressures and diversify solutions. The use of coevolution addresses three fundamental challenges faced in past fitness approximation research: 1) the model learning investment; 2) the level of approximation of the model; and 3) the loss of accuracy. We discuss applications of this approach and demonstrate its impact on the symbolic regression problem. We show that coevolved predictors scale favorably with problem complexity on a series of randomly generated test problems. Finally, we present additional empirical results that demonstrate that fitness prediction can also reduce solution bloat and find solutions more reliably.      

基于Logistic测试覆盖率函数的软件可靠性建模研究

软件测试覆盖率是测试充分性和测试效率的有效度量指标,其与软件可靠性以及缺陷覆盖情况之间有着一定的相关关系,并且结合测试覆盖率信息的软件可靠性模型的评估和预计效果将会得到有效改进.在实际测试过程中,由于软件结构特征及学习因素的综合影响,测试覆盖率可能会呈现出一种先增后减的趋势,Logistic函数恰好非常适合描述这类S形变化趋势,且结构简单,具有较好的灵活性与适应性.因此,针对基于Logistic函数的测试覆盖率函数以及软件可靠性建模等问题展开研究.首先提出基于Logistic函数的测试覆盖率函数;在该函数的基础上,提出基于Logistic测试覆盖函数的缺陷预计模型;然后,将NHPP可靠性模型的建模过程与Logistic测试覆盖函数相结合,提出一种新的者虑测试覆盖率的软件可靠性增长模型.实例验证结果表明:与若干已有的同类研究成果相比,提出的基于Logistic函数的测试覆盖率函数、缺陷预计模型以及软件可靠性增长模型有效地提高了函数或模型对数据的拟和精度,且具有较好的适用性.     

Model identification using tie line power and frequency measurements

The identification of dynamic models which relate power and frequency deviations on a tie line of a power system is investigated. The identification problem is posed and three identification algorithms are presented which produce least squares models with different structural properties. Model order is determined by applying residual and system structure tests to a sequence of models of increasing order. These tests indicate the model order for both equivalent realizations and predictive models. Equivalent realizations are identified on one data set and then their performance as a dynamic equivalent is evaluated on a second data set. These equivalent realizations are also used to predict frequency in an iterative frequency prediction algorithm. Predictive models are also identified and their performance as frequency predictors is evaluated using a direct prediction algorithm. The identification of dynamic equivalents provides information about the structural properties of power systems. The use of dynamic equivalents and predictive models for frequency prediction indicates the tradeoff in accuracy vs the prediction interval which can be obtained using these least squares algorithms and the measurement device presently available.     

基于极限学习机的组合软件可靠性模型研究

针对单一软件可靠性模型适应性不强和数据驱动模型稳定性较差的问题,本文选取3种典型软件可靠性模型作为基模型,利用极限学习机对基模型的预测结果进行加权优化,得到组合软件可靠性模型,实现经典软件可靠性模型和人工智能算法的有机结合。通过对3组失效数据进行仿真实验,并与单一模型、基于其他神经网络算法的组合模型以及数据驱动模型的预测结果进行对比,验证了本文模型能够有效地提升预测精度和模型的适应性。     

多路径覆盖测试数据生成适应度函数设计方法

综合考虑影响适应度函数设计的因素,提出一种基于层次分析法的适应度函数设计方法。该方法首先将影响路径之间相似度的因素归结为三要素,并建立层次分析模型。根据不同因素对路径间相似度的作用重要程度不同,建立因素之间两两比较的判断矩阵,确定每个因素的权重系数,进而构造适应度函数。最后,将该方法用于基于遗传算法的多路径覆盖的测试数据生成。实验结果表明,对于解决多路径覆盖的测试数据生成问题,与已有方法相比,该方法具有较好的优越性。     

Probabilistic Boolean network modeling of an industrial machine

Theoretical modeling of manufacturing processes assists the design of new systems for predictions of future behavior, identifies improvement areas, and evaluates changes to existing systems. A novel approach is proposed to model industrial machines using probabilistic Boolean networks (PBNs) to study the relationship between machine components, their reliability and function. Once a machine is modeled as a PBN, through identification of regulatory nodes, predictors and selection probabilities, simulation and property verification are used to verify model correctness and behavior. Using real machine data, model parameters are estimated and a PBN is built to describe the machine, and formulate valid predictions about probability of failure through time. Two models were established: one with non-deterministic inputs (proposed), another with components’ MTBFs inputs. Simulations were used to generate data required to conduct inferential statistical tests to determine the level of correspondence between predictions and real machine data. An ANOVA test shows no difference between expected and observed values of the two models (p value = 0.208). A two-sample T test demonstrates the proposed model provides values closer to expected values; consequently, it can model observable phenomena (p value \(=\) 0.000). Simulations are used to generate data required to conduct inferential statistical tests to determine the level of correspondence between model prediction and real machine data. This research demonstrates that using PBNs to model manufacturing systems provides a new mechanism for the study and prediction of their future behavior at the design phase, assess future performance and identify areas to improve design reliability and system resilience.     

基于泛函网络的软件可靠性多模型综合预测方法

将泛函网络引入软件可靠性预测,利用其比神经网络更好的解释性及其他性能,提出了基于泛函网络的软件可靠性多模型综合预测方法。首先阐述了泛函网络的结构和学习过程,然后将多个单一模型的预测值作为泛函网络的输入,将实际值作为输出,建立泛函网络结构,给出了泛函网络的学习算法,制定了3种训练策略,并进行了实验分析。实验结果表明:在第三种训练策略下,基于泛函网络的软件可靠性多模型综合预测方法有较高的预测精度,其预测效果比单个模型和Lyu提出的线性综合模型都好。     

Discrete-time point processes in urban traffic queue estimation

This research was motivated by the belief that it is possible to develop improved algorithms for the computer control of urban traffic. Previous research suggested that the computer software, and especially the filtering and prediction algorithms, is the limiting factor in computerized traffic control. Since the modern approach to filtering and prediction begins with the development of models for the generation of the data and since these models are also useful in the control problem, this paper deals with the modeling of traffic queues and filtering and prediction. It is shown that the data received from vehicle detectors is a discrete-time point process. The formation and dispersion of queues at a traffic signal is then modeled by a discrete-time time-varying Markov chain which is related to the observation point process. Three such models of increasing complexity are given. Recent results in the theory of point-process filtering and prediction are then used to derive the nonlinear minimum error variance filters/predictors corresponding to these models. It is then shown that these optimal estimators are computationally feasible in a micro-processor. All three algorithms were tested against the UTCS-1 traffic simulator and, in one case, against an algorithm in current use called ASCOT. Some results of these tests are shown. They indicate good performance in every case and better performance than ASCOT in the comparable case.     

Effects of computer-based test administrations on test anxiety and performance

This study examined the degree to which computer and test anxiety had a predictive role in performance across three computer-administered placement tests (math, reading, written English). Subjects (72 college undergraduates at a Midwestern university) were measured with the Computer Anxiety Rating Scale, the Test Anxiety Inventory, the Myers-Briggs Type Indicator (MBTI), and the three achievement areas. Age and gender were added to the prediction model for completeness. Results showed that age and test anxiety were significant predictors for math performance, with lower values on the two variables associated with better performance. When reading was the outcome variable, age and computer anxiety were statistically significant performance predictors, with older readers faring better and less anxious individuals achieving higher scores. No predictors were statistically significant for the written English essay. The hypothesis that “thinkers” and “intuitives” on the MBTI would have lower anxiety scores was only partially confirmed. Thinkers had lower anxiety scores than “feelers,” but there was no difference on the intuitive-sensor dimension. The results suggest that much of what is considered computer anxiety may in fact be a manifestation of test anxiety. It is possible that, by giving examinees more perceived control, anxiety levels can be reduced.     

Development of a physical employment standard for a branch of the UK military

Abstract

A Physical Employment Standard (PES) was developed for the British Royal Air Force Regiment (RAF Regt). Twenty-nine RAF Regt personnel completed eight critical tasks wearing Combat Equipment Fighting Order (31.5?kg) while being monitored for physical and perceptual effort. A PES was developed using task simulations, measured on 61 incumbents. The resultant PES consists of: 1) a battlefield test involving task simulations: single lift and point-of-entry (psss/fail); timed elements (react to effective enemy fire and crawl) set at 95th performance percentile; casualty evacuation (CASEVAC) casualty drag and CASEVAC simulated stretcher carry completed without stopping. 2) a Multi Stage Fitness Test level 9.10 to assess aerobic fitness to complete a tactical advance to battle. The task-based PES should ensure RAF Regt personnel have a baseline level of fitness to perform and withstand the physical demands of critical tasks to at least a minimum acceptable standard.

Practitioner summary: A Physical Employment Standard (PES) was developed for the British RAF Regiment by measuring the physiological demands of critical tasks on a representative cohort of incumbent personnel. A task-based PES should ensure that only those candidates, irrespective of gender, race or disability, with the necessary physical attributes to succeed in training and beyond, are selected.     

可靠性模型中故障检测率研究述评

故障检测率FDR（Fault Detection Rate）是可靠性研究的关键要素,对于测试环境构建、故障检测效率提升、可靠性建模和可靠性增长具有重要作用,对于提高系统可靠性与确定发布时间具有重要现实意义.首先,对基于NHPP（Non-Homogeneous Poisson Process,非齐次泊松过程）类的软件可靠性增长模型SRGM（Software Reliability Growth Mode）进行概述,给出了建模本质、功用与流程.基于此,引出可靠性建模与研究中的关键参数——FDR,给出定义,对测试环境描述能力进行分析,展示不同模型的差异.着重剖析了FDR与失效强度、冒险率（风险率）的区别,得出三者之间的关联性表述.全面梳理了FDR的大类模型,分别从测试覆盖函数视角、直接设定角度、测试工作量函数参与构成方式三个方面进行剖析,继而提出统一的FDR相关的可靠性模型.考虑到对真实测试环境描述能力需要,建立不完美排错框架模型,衍生出不完美排错下多个不同FDR参与的可靠性增长模型.进一步,在12个真实描述应用场景与公开发表的失效数据集上进行实验,验证不同FDR模型相关的可靠性模型效用,对差异性进行分析与讨论.结果表明,FDR模型自身的性能可以支撑可靠性模型性能的提升.最后,指出了未来研究趋势和需要解决的问题.     

工业控制系统可靠性设计与测试评估技术研究

目前我国的工业控制系统发展迅速,但其可信性较低,为解决这一问题,我们开展了工业控制系统的可靠性设计和测试评估技术研究。研究形成了工业控制系统可靠性设计规范和工业控制系统可靠性测试评估规范,另外还生成了工业控制系统可靠性设计和测试评估的软件操作平台。     

Estimation of individual prediction reliability using the local sensitivity analysis

For a given prediction model, some predictions may be reliable while others may be unreliable. The average accuracy of the system cannot provide the reliability estimate for a single particular prediction. The measure of individual prediction reliability can be important information in risk-sensitive applications of machine learning (e.g. medicine, engineering, business). We define empirical measures for estimation of prediction accuracy in regression. Presented measures are based on sensitivity analysis of regression models. They estimate reliability for each individual regression prediction in contrast to the average prediction reliability of the given regression model. We study the empirical sensitivity properties of five regression models (linear regression, locally weighted regression, regression trees, neural networks, and support vector machines) and the relation between reliability measures and distribution of learning examples with prediction errors for all five regression models. We show that the suggested methodology is appropriate only for the three studied models: regression trees, neural networks, and support vector machines, and test the proposed estimates with these three models. The results of our experiments on 48 data sets indicate significant correlations of the proposed measures with the prediction error.     

Prediction of drill flank wear using ensemble of co-evolutionary particle swarm optimization based-selective neural network ensembles

Flank wear prediction plays an important role in achieving improved productivity and better quality of the product. This study presents an effective co-evolutionary particle swarm optimization-based selective neural network ensembles (E-CPSOSEN) enabled tool wear prediction model for flank wear prediction in drilling operations. The E-CPSOSEN algorithm utilized two populations of particle swarm optimizations (PSOs) that are co-evolved simultaneously, one discrete particle swarm optimizations for evolving the binary selection vector, and the other continuous particle swarm optimizations for evolving the real weight vector. The two PSOs interact with each other through the fitness evaluation. The E-CPSOSEN algorithm is first tested on four benchmark problems taken from the literature. Upon achieving good results for test cases, the E-CPSOSEN enabled tool wear prediction model was employed to three illustrative case studies of flank wear prediction in drilling operations. Significant improvement is also obtained in comparison to the results already reported in literatures, which further reveals that the E-CPSOSEN enabled tool wear prediction model has more wonderful prediction performance than conventional single ANN-based models in predicting the flank wear in drilling operations. Moreover, an investigation was also conducted to identity the effects of the major parameters of the E-CPSOSEN algorithm upon its prediction performance. From the given results, the proposed enabled tool wear prediction model may be a promising tool for the accurate and automatic prediction of flank wear in drilling operations.     

Rock strength assessment based on regression tree technique

Uniaxial compressive strength (UCS) is one of the most important parameters for investigation of rock behaviour in civil and mining engineering applications. The direct method to determine UCS is time consuming and expensive in the laboratory. Therefore, indirect estimation of UCS values using other rock index tests is of interest. In this study, extensive laboratory tests including density test, Schmidt hammer test, point load strength test and UCS test were conducted on 106 samples of sandstone which were taken from three sites in Malaysia. Based on the laboratory results, some new equations with acceptable reliability were developed to predict UCS using simple regression analysis. Additionally, results of simple regression analysis show that there is a need to propose UCS predictive models by multiple inputs. Therefore, considering the same laboratory results, multiple regression (MR) and regression tree (RT) models were also performed. To evaluate performance prediction of the developed models, several performance indices, i.e. coefficient of determination (R ²), variance account for and root mean squared error were examined. The results indicated that the RT model can predict UCS with higher performance capacity compared to MR technique. R ² values of 0.857 and 0.801 for training and testing datasets, respectively, suggests the superiority of the RT model in predicting UCS, while these values are obtained as 0.754 and 0.770 for MR model, respectively.     

面向节点-节点的进化测试改进方法

传统方法不能对进化测试中所有面向节点-节点的测试类型都构造具有良好导向的适值函数。针对该问题,基于面向节点-节点进化测试系统模型,不考虑节点的执行顺序和控制流关系,从节点的独立性出发,提出一种改进的适值函数计算方法。实验结果表明,对离散节点之间没有数据依赖关系的覆盖准则,该方法代价较小、运行稳定。