This paper presents a systematic and general method for Lamb wave-based crack size quantification using finite element simulations and Bayesian updating. The method consists of construction of a baseline quantification model using finite element simulation data and Bayesian updating with limited Lamb wave data from target structure. The baseline model correlates two proposed damage sensitive features, namely the normalized amplitude and phase change, with the crack length through a response surface model. The two damage sensitive features are extracted from the first received S₀ mode wave package. The model parameters of the baseline model are estimated using finite element simulation data. To account for uncertainties from numerical modeling, geometry, material and manufacturing between the baseline model and the target model, Bayesian method is employed to update the baseline model with a few measurements acquired from the actual target structure. A rigorous validation is made using in-situ fatigue testing and Lamb wave data from coupon specimens and realistic lap-joint components. The effectiveness and accuracy of the proposed method is demonstrated under different loading and damage conditions.

Software defect prediction (SDP) is an effective technique to lower software module testing costs. However, the imbalanced distribution almost exists in all SDP datasets and restricts the accuracy of defect prediction. In order to balance the data distribution reasonably, we propose a novel resampling method LIMCR on the basis of Naïve Bayes to optimize and improve the SDP performance. The main idea of LIMCR is to remove less-informative majorities for rebalancing the data distribution after evaluating the degree of being informative for every sample from the majority class. We employ 29 SDP datasets from the PROMISE and NASA dataset and divide them into two parts, the small sample size (the amount of data is smaller than 1100) and the large sample size (larger than 1100). Then we conduct experiments by comparing the matching of classifiers and imbalance learning methods on small datasets and large datasets, respectively. The results show the effectiveness of LIMCR, and LIMCR+GNB performs better than other methods on small datasets while not brilliant on large datasets.

Integrated modular avionics (IMA) systems present many advantages. However, the resource sharing mechanism also brings a series of system problems, including the frequency of fault propagation and the difficulties of system design verification. The traditional analysis approaches for system designers have limits to analyze dynamic faults which are caused by unreasonable designs. These dynamic faults come up with component fault states, component state correlation, and system dynamic behaviors. In this paper, a new model-based dynamic analysis method for state correlation with IMA fault recovery is proposed, which helps to check system states and verify system designs by means of analyzing the dynamic behaviors of systems in a new view of systems' correlated states. A colored generalized stochastic Petri net (CGSPN) provides advantages to system modeling and simulation, but there are some difficulties for modeling component state correlations and system dynamic behaviors in detail on the IMA system. We make an improvement on CGSPN for modeling IMA by adding an element and changing fairing rules. In addition, multiconstraint specified to solve the configuration satisfying problem for IMA is built into the model. Afterward, according to results of model simulation, system dynamic faults are analyzed and system designs are checked, which will help to guide the system designers to adjust system architecture at the early stage of system development. Finally, a case study is given for demonstrating how to apply this new method.

Test case prioritization (TCP) has been considerably utilized to arrange the implementation order of test cases, which contributes to improve the efficiency and resource allocation of software regression testing. Traditional coverage-based TCP techniques, such as statement-level, method/function-level and class-level, only leverages program code coverage to prioritize test cases without considering the probable distribution of defects. However, software defect data tends to be imbalanced following Pareto principle. Instinctively, the more vulnerable the code covered by the test case is, the higher the priority it is. Besides, statement-level coverage is a more fine-grained method than function-level coverage or class-level coverage, which can more accurately formulate test strategies. Therefore, we present a test case prioritization approach based on statement software defect prediction to tame the limitations of current coverage-based techniques in this paper. Statement metrics in the source code are extracted and data pre-processing is implemented to train the defect predictor. And then the defect detection rate of test cases is calculated by combining the prioritization strategy and prediction results. Finally, the prioritization performance is evaluated in terms of average percentage faults detected in four open source datasets. We comprehensively compare the performance of the proposed method under different prioritization strategies and predictors. The experimental results show it is a promising technique to improve the prevailing coverage-based TCP methods by incorporating statement-level defect-proneness. Moreover, it is also concluded that the performance of the additional strategy is better than that of max and total, and the choice of the defect predictor affects the efficiency of the strategy.

Heterogeneous defect prediction (HDP) predicts defects for the current project (target) using heterogeneous data from other projects (source), which can improve software quality effectively. The ability to reduce the distribution divergence between the source and target project is crucial to the performance of HDP. Existing HDP methods address this issue only by using feature adaptation techniques, i.e., feature matching and feature transformation. However, without considering instance adaptability differences, they cannot fully exploit the potential of instance adaptation and thus cannot make the source and target distribution completely matched. To address this problem, we propose a novel HDP approach called joint instance and feature adaptation (JIFA) combining both instance adaptation and feature adaptation to narrow the gap between projects. Making joint use of both adaptation techniques, JIFA not only further reduces the distribution discrepancy but also enhances its robustness to abnormal outliers. Besides, discriminant information is also preserved in JIFA so that a better classification boundary can be obtained. Particularly, JIFA applies to both heterogeneous and homogeneous cross-project defect prediction (CPDP) tasks. The experimental results from 22 projects indicate that JIFA outperforms a range of advanced heterogeneous and homogeneous CPDP approaches.