With the growing concern for autonomous vehicle (AV) reliability, various statistical metrics have been developed to measure their long-term and average behaviors. However, these metrics overlook the characteristics during the phased-mission operations of AVs. This article proposes a novel method to dynamically assess the mission reliability of AV systems. We first establish a dedicated mission reliability metric specifically tailored for AV applications. A probabilistic assessment model is then developed to consider time-varying mission demands, mission criticality, environmental dependence, and measurement noises. Integrating the local linear regression model and sample average approximation approach, the distribution of mission performance is analyzed to address nonlinear environmental dependencies. The theoretical proof for the finite-sample performance guarantee of our model is rigorously established. Numerical simulations demonstrate the method’s superior performance in finite-sample scenarios over Monte Carlo approaches. A real-world case study on lateral vehicle control further confirms the effectiveness of the method.

The extreme requirements for high reliability and low latency in the upcoming Sixth Generation (6G) wireless networks are challenging the design of multi-hop wireless transport networks. Inspired by the advent of the virtualization concept in the wireless networks design andopennessparadigm as fostered by the Open-Radio Access Network (O-RAN) Alliance, we target a revolutionary resource allocation scheme to improve the overall transmission efficiency. In this paper, we investigate the problem of automatic repeat request (ARQ) in multi-hop decode-and-forward (DF) relaying in the finite blocklength (FBL) regime, and propose a dynamic scheme of multi-hop hybrid ARQ (HARQ), which maximizes the end-to-end (E2E) communication reliability in the wireless transport network.We also propose an integer dynamic programming (DP) algorithm to efficiently solve the optimal Dynamic Multi-Hop HARQ (DMH-HARQ) strategy. Constrained within a certain time frame to accomplish E2E transmission, our proposed approach is proven to outperform the conventional listening-based cooperative ARQ, as well as any static HARQ strategy, regarding the E2E reliability. It is applicable without dependence on special delay constraint, and is particularly competitive for long-distance transport network with many hops.

As concerns about the reliability of autonomous vehicles (AVs) continue to rise, various statistical metrics have been developed to evaluate their long-term and average failure behaviors. However, these metrics often overlook the unique characteristics of AVs’ specific mission performance. The AV system, equipped with an intricate computing system, is significantly influenced by end-to-end latencies, spanning from sensors to control signals. Previous research has focused on the impact of latencies within individual subsystems, particularly the control subsystem, without examining these impacts at the broader computing system level. Additionally, the state dependence of these latencies remains unexplored. This paper introduces a mission reliability assessment method specifically designed for AVs, considering the state dependence on end-to-end latencies. We propose a dedicated metric for mission reliability in AV systems, tailored to capture the features of end-to-end latencies. We apply the hidden Markov model to analyze the transition process of end-to-end latencies and estimate the mission reliability. The effectiveness of our method is validated through two numerical simulation cases, demonstrating its capacity for real-time evaluation and offering significant benefits for the online management and operational maintenance of AVs.

Autonomous Driving Systems (ADS) are considered safety-critical, as even a minor fault may lead to catastrophic consequences. To evaluate their reliability and robustness under failure conditions, Fault Injection (FI) techniques have been widely adopted. Most existing FI methods employ data-driven approaches, such as surrogate modeling and reinforcement learning, to generate test cases. While these techniques have shown promise, they often incur substantial costs in terms of data collection and training time. Moreover, their performance is highly sensitive to the quality and quantity of training data, which can limit their applicability in diverse or unseen scenarios. In this paper, we propose LOFT, an efficient multi-objective search-based FI testing framework that leverages Large Language Models (LLMs) to identify diverse and realistic critical faults. To accommodate the structured and non-linguistic nature of raw simulation data, LOFT adopts a two-stage LLM-based fault injection pipeline. In the first stage, an LLM converts singleframe simulation data into natural language descriptions and suggests appropriate fault types. In the second stage, a separate LLM examines the broader scenario context to determine the optimal time window for fault injection. The outputs from the two LLMs are then used to initialize and guide a multi-objective search procedure aiming at discovering a diverse set of critical faults. We implement LOFT and evaluate on an ADS provided by our industrial partner. Experimental results show that, compared with two baseline approaches, LOFT detects over $90 \%$ more critical faults and identifies an average of 2.2 additional fault types within an equivalent number of simulations.

With the increasing demand for high-quality telecommunication services, cellular KPI prediction becomes crucial for telecommunication network monitoring and management. In this work, we propose a novel framework for cellular KPI prediction, which considers its distribution discrepancy under different network operation scenarios. In particular, three specific predictors for normal, target alarm, and neighbor alarm scenarios are proposed based on spatiotemporal graph neural networks and unified through transfer learning. Temporal convolution and attention mechanism are embedded to model the impact of anomalies on KPIs and its propagation across neighboring cells according to the cellular network topology. An experiment on a real cellular KPI dataset shows the effectiveness of the proposed method compared to the state-of-the-arts. Note to Practitioners—Cellular network KPI prediction under scenarios of network alarms is crucial to evaluate the impact of alarms on network services and guides cellular network maintenance policies. This problem is similar to a general multivariate time series prediction problem with data multimodality. However, the first challenge in our case is that, under different scenarios, i.e., normal, target alarm, and neighbor alarm, the effective information and spatiotemporal dependencies among KPIs are different. The second challenge is the imbalanced or sparse sample size for specific scenarios, deteriorating the model performance. This paper proposes a cellular KPI prediction framework consisting of three scenario-specific predictors with similar but different modules to process different scenario-specific data. To address the dataset imbalance across scenarios, we adopt a transfer learning strategy to unify the training and prediction of three predictors. The experiment results on a real cellular KPI dataset demonstrate that the proposed framework is more feasible and effective than the state-of-the-art models for multivariate time series prediction. Future research can consider developing maintenance policies such that the cost caused by abnormal KPIs can be minimized.

With the rapid advancement of sensing technology, some critical components within engineering systems are equipped with sensors to collect condition monitoring (CM) signals. Such systems are referred to as partially monitored systems because only selected components are monitored. However, the method to integrate real-time component-level CM signals into reliability assessments of these systems remains unexplored. This study introduces a novel reliability assessment method designed to address the challenges of evaluating the reliability of partially monitored systems, particularly considering the highly nonstationary nature of CM signals and their dependence on the component states. A multistate degradation hidden Markov model with time-varying parameters (DHMM-TVP) is developed to better handle the nonstationary and nonlinear nature of CM signals. The expectation-maximization (EM) algorithm is adapted to estimate the unknown parameters within the DHMM-TVP framework. Furthermore, leveraging DHMM-TVP in combination with a functional kernel regression model, a generalized reliability assessment method is proposed, specifically tailored for cases where the system reliability structure is unknown or only partially known. A numerical simulation study and two case studies were conducted to validate the proposed reliability assessment approach. The component-level validation was performed using an experimental bearing accelerated degradation testing dataset, while the system-level verification employed aircraft turbofan engine datasets from the NASA prognostics data repository, collectively demonstrating the effectiveness of the proposed method.

Addressing the issue of sample imbalance in vision-based anomaly detection tasks remains a critical focus. This work proposes a novel hybrid method that integrates learning-based multidimensional feature extraction with a Novel Adaptive Undersampling Framework (NAUF) for image-based anomaly detection, particularly when defect samples are extremely scarce. First, the proposed method extracts image features using the backbone of a pretrained deep learning network. Next, the undersampling technique NAUF is applied to these extracted features, balancing the highly imbalanced image samples while preserving essential information. Finally, the images are classified using multiple base classifiers within an ensemble learning framework. On a dataset of defects for the inner surfaces of high-pressure hydrogen storage tanks, the proposed method significantly improved the key performance metrics (AUCPRC) of KNN, Decision Tree, and Random Forest by 14.23%, 11.67%, and 7.52% respectively, while also increasing their inference speeds by 60.43%, 97.34%, and 86.36%, en-hancing the practical application value of these base classifiers. This work offers valuable insights and potential applications for improving quality control in automated manufacturing and other industrial settings where data imbalance is a common challenge.

Government bodies, companies, and social agencies typically act as defenders and apply the game theory method to protect their assets targeted by malicious attackers. However, for developed transregional or even transnational systems, the upper-level administrator fails to coordinate the lower-level agents (sub-system defenders) due to the organizational and beneficial boundaries. This work focuses on the anarchy phenomenon of such decentralized non-cooperative multiagent systems, and proposes a non-cooperative multi-defender Stackelberg defense resources allocation game model to analyze the attack defense confrontation process. We also discuss homogeneous systems analytically and explore general systems numerically to reveal the relationships between the anarchy phenomenon and system parameters, which extends the generality of such problems and provides the upper-level administrator with useful information on constructing a decentralized multiagent system. Besides, to mitigate the anarchy phenomenon, this work first designs a risk-sharing mechanism to support the upper-level administrator in coordinating non-cooperative defenders from an individual-optimal defense strategy to a global-optimal one. The proposed risk-sharing mechanism is established, and the anarchy phenomenon is analyzed to the target system managed by multiple non-cooperative defenders. Simulations on the local metering system also show small investment in the risk-sharing mechanism could reduce considerable total social welfare loss.Note to Practitioners—This paper is motivated by the practical demand of protecting decentralized non-cooperative multiagent systems consisting of several beneficial-independent agents (sub-system defenders) and one weak administrator, such as cross-region power grids, and cross-region water distribution systems. Different from previous works allocating resources global-optimally, this study focuses on excessive or insufficient resources allocation due to the anarchy phenomenon. We then design a risk-sharing mechanism for the administrator of the decentralized non-cooperative multiagent system. Based on this, subsystem system operators would follow the instruction from the system-level administrator instead of implementing a local-optimal defense strategy. The proposed model and method are demonstrated by numerical experiments on the local metering system, and can be applied to many decentralized multiagent systems. For example, non-cooperative power companies would follow the instruction on global-optimal defense resources allocation on smart meters from the upper-level administrator.

We propose a novel method for weakly-supervised anomaly detection, where a limited number of labeled normal samples and a sufficient number of unlabeled samples are available for modeling. In particular, we seamlessly integrate label propagation with manifold graph learning into a support vector data description model. Consequently, the estimated manifold graph as well as its parameters will be adaptive to label propagation and benefit the anomaly detection performance. It is superior to most graph-based models that perform manifold graph learning separately by an independent step before label propagation. Theoretically, we derive a stability analysis based on the Rademacher complexity. Further, the effectiveness of the proposed method is demonstrated through several benchmark data sets and a real example of fault detection for high-speed train wheels.Note to Practitioners—This article provides a weakly-supervised anomaly detection method and addresses the challenge of insufficient normal samples for training. The proposed method integrates the adaptive embedded label propagation with adaptive manifold graph learning into a support vector data description model to additionally exploit the intrinsic data distribution information of the unlabeled data in the model formulation. It ensures that the results are jointly optimal for manifold representation and anomaly detection such that the detection accuracy is improved.

The challenges of class-imbalance and partially unknown training labels often arise in fault detection tasks. When these two problems occur simultaneously, existing imbalanced classification methods cannot be directly used due to the absence of the label, and the class-imbalance would lead to severe bias prediction. In this study, we proposed a novel deep imbalance separation network (deepImSN) framework that is capable of dealing with fault detection problems with both class imbalance and partially unknown labels. This framework integrates the one-class learning concept into the positive-unlabeled (PU) learning theory for the first time. It alleviates the bias of the class-imbalance while making full use of the limited label information in the PU set to optimize the feature space and guide model training. The proposed deepImSN is designed to be used in different scenarios. It can accurately complete the fault detection task whether only part of fault samples or normal samples are labeled, and the class-prior is known or unknown. Experimental results on real-world problems, such as high-speed rail wheels fault inspection and wafer map fault detection, demonstrate that deepImSN outperforms existing methods in various experimental conditions.

Industrial visual monitoring (IVM) is crucial in enhancing the reliability and efficiency of manufacturing processes. Recently, large vision-language models (LVLMs) have demonstrated remarkable semantic understanding and natural language interaction capabilities, which provide a novel solution to IVM. However, LVLMs pretrained on common domains lack specific knowledge for IVM scenarios, causing insufficient adaptation to industrial image patterns and specialized textual corpora. In this article, we deeply studied the adaptation of LVLMs to IVM and proposed DefectGLM. First, we proposed the first large-scale multimodal wafer dataset as a reliable data basis for model domain generalization. Second, this model employs low-rank adaptation–based contrast visual adaptation to align with industrial image patterns and utilizes vision-language instruction tuning for professional knowledge alignment. DefectGLM is the first large-model-based wafer image recognition model, and can accurately identify 36 types of wafer defects and provide appropriate text descriptions. DefectGLM provides a new solution for the development of industrial large models.

The wheel wear status of high-speed trains (HSTs) is an essential indicator of their safety and reliability. However, due to the time-varying operating state of HSTs, noisy and complex non-stationary signals are collected. This makes it difficult for data-driven algorithms to learn valuable discriminative features from data. Therefore, this inspired us to introduce signal analysis methods with clear physical meaning to improve the interpretability and performance of prediction models. This paper proposes a novel multi-layer wavelet integrated convolutional neural network (MWI-Net) for predicting HST wheel-wear. Specifically, discrete wavelet transform (DWT) extends the feature learning space of CNN from the time domain to the wavelet domain, thereby capturing the frequency features that are difficult to learn in the time domain. As a remarkable information space, the DWT can effectively alleviate the frequency aliasing problem, enabling MWI-Net to distinguish valuable frequency information from complex signals. In particular, the proposed dynamic frequency aggregation mechanism endows MWI-Net with excellent frequency analysis and feature selection capabilities. Experiments on the real operation dataset of CRH1A HSTs show that MWI-Net accurately predicts the wheel wear curves, which is more competitive than existing deep learning methods. Furthermore, we demonstrate the feature learning mechanism inside MWI-Net through visual analysis and illustrate how it optimizes and extracts valuable features layer by layer.

The wheels are among the most critical components which largely influence the safe operation of high-speed trains. The existing research in reliability modeling typically assumes the wheel degradation to be a one-dimensional degradation process. This could incur a deficiency in practice, as the wheel degradation is in fact the superposition of multiple complex degradation processes, involving wheel tread wear and wheel polygonal wear. Random shocks also contribute to the wheel degradation. Moreover, these processes are correlated with each other. To fully consider these factors, this article proposes a multistate model for multidimensional degradations. The piecewise-deterministic Markov process (PDMP) model is applied to describe the mutual dependencies between random shocks and multiple degradations. Conventionally, the parameters of PDMP are set by experts’ experience. This article investigates maximum likelihood estimation to estimate the model parameters. Finally, the Monte Carlo simulation algorithm is proposed to evaluate the high-speed train wheel's reliability. Numerical experiments were conducted to validate the proposed method on high-speed train wheels subject to tread wear, polygonal wear, and wheel-rail impacts, which show that dependencies among multidimensional degradation processes and random shocks will largely affect the reliability of the wheels. The application to high-speed train wheels shows the effectiveness of the proposed model.

High-level maintenance is one of the key practices to ensure the safe operation of high-speed trains (HSTs). However, it is usually scheduled manually, and there has been a lack of systematic study on how to optimize maintenance scheduling considering a dynamic time interval. This article discusses the state-of-the-art high-level maintenance scheduling models of HSTs. The one-time maintenance scheduling problem, where the maintenance action occurs at most once in the planning horizon, is a key issue for HSTs’ high-level maintenance. We construct a more efficient mathematical model and propose an exact solution for the one-time scheduling model (OSM). However, HSTs might be maintained several times within the planning horizon to achieve a more favorable performance in real-world practice. Therefore, we extend the OSM to the periodic scheduling model (PSM), which aims to accommodate several maintenance actions and could be valuable to obtain a global optimal plan. Both models are mixed 0–1 linear integer models and can be solved by commercial solvers. The effectiveness and comparisons of the proposed models are illustrated by a real-case study using the maintenance records of 100 HSTs in China. The results can serve as the basis for the HSTs’ high-level maintenance scheduling under either a periodic case or a one-time case.

Machine intelligence fault prediction (MIFP) is crucial for ensuring complex systems’ safe and reliable operation. While deep learning has become the mainstream tool for MIFP due to its excellent learning abilities, its interpretability is limited, and it struggles to learn frequencies, making it challenging to understand the physical knowledge of signals at the frequency level. Therefore, this article proposes a physically interpretable wavelet-guided network (WaveGNet) with deep frequency separation for MIFP, inspired by the sound theoretical basis and physical meaning of discrete wavelet transform (DWT). WaveGNet expands the feature learning space of CNN into the frequency domain, allowing for a better understanding of the physical insights behind the frequency level. Specifically, WaveGNet involves a derivable and learnable frequency learning layer (FL-Layer) consisting of a wavelet-driven frequency decomposition module and a convolution-driven feature learning module. Multiple DWT-driven FL-Layers are used in WaveGNet to achieve deep frequency decomposition and multiresolution frequency feature learning in a coarse-to-fine manner. The effectiveness of WaveGNet was evaluated in real high-speed train wheel wear monitoring and high-speed aviation bearing fault diagnosis cases. Experimental results showed that WaveGNet outperforms cutting-edge deep learning algorithms and has excellent fault diagnosis and prediction abilities. Furthermore, an in-depth analysis of the learning mechanism of wavelet-driven CNN from the frequency domain perspective was conducted.

Autonomous Driving Systems (ADS) are safety-critical and require comprehensive testing before their deployment on public roads. Most existing testing approaches consist in generating scenarios that vary the behaviors of dynamic objects, while leaving a predefined road environment unchanged. Consequently, these approaches overlook the influence of different road structures on ADS safety, e.g., collisions can happen more frequently than usual on a merging road, because of the specific road structure. In this paper, we propose EvoScenario, a novel approach that integrates road structures into the generation of critical scenarios for exposing safety risks of ADS. Specifically, EvoScenario models a driving road as a sequence of road segments characterized in different aspects, such as their shapes and widths. Then, a test case is defined by concatenating the sequence of road segments and the sequence of dynamic object maneuvers. Inspired by EvoSuite that generates sequential method calls for Java unit testing, EvoScenario leverages the sequential models of test cases and constructs a multi-objective optimization framework to search for critical scenarios. We implement and demonstrate EvoScenario on an ADS provided by our industrial partner. Evaluation results show that EvoScenario can identify 6 types of safety violations, and outperform existing baseline testing approaches.

Autonomous driving systems (ADS) are safety-critical systems that require thorough testing to ensure their safety. Current testing methods for ADS primarily focus on finding crash scenarios involving ADS. However, most of these scenarios are unavoidable by ADS, such as collisions caused by the reckless behavior of other vehicles. To address this limitation, we propose CollVer, a framework designed to generate and identify scenarios in which ADS violate driving rules. Specifically, CollVer utilizes multi-modal technology by taking the violation scenario and the corresponding accident description as inputs to judge whether the accident can be attributed to the ADS. Moreover, CollVer introduces a metric called collision position coverage (CPC), to quantify and guide the selection of test cases. Finally, CollVer integrates the multi-modal model and the CPC metric into a multi-objective genetic algorithm to explore more diverse and challenging scenarios. We evaluate CollVer on an industrial-grade ADS, Baidu Apollo, and experimental results show that CollVer can identify 10 distinct types of safety violations, with 4 of them resulting from ADS violating driving rules.

In practical industrial applications, the need for a large number of accurately labeled training samples is a significant challenge for fault detection tasks. However, labeling all training samples is expensive and prone to labeling errors, especially for early fault detection of wind turbine blades. This article proposes a labeling bias hypothesis. Assuming the labeler only needs to label parts of normal samples that are easy to judge, we design a probability ratio least-squares importance fitting (PRL-SIF) method based on variable homogeneity. Unlike other state-of-the-art positive unlabeled learning methods, PRL-SIF does not require knowledge of the class priors to achieve training. Furthermore, to better handle the multidimensional time-series data of wind turbines, we provide a data preprocessing method based on functional analysis to achieve time series feature extraction and dimensionality reduction. The effectiveness and robustness of the proposed method are verified on 23 real-world wind turbine datasets. Experimental results show that the proposed method can achieve nearly 90% accuracy while only labeling 20% of normal samples.

We propose a probabilistic model for clustering spatially correlated functional data with multiple scalar covariates. The motivating application is to partition the 29 provinces of the Chinese mainland into a few groups characterized by the epidemic severity of COVID-19, while the spatial dependence and effects of risk factors are considered. It can be regarded as an extension of mixture models, which allows different subsets of covariates to influence the component weights and the component densities by modeling the parameters of the mixture as functions of the covariates. In this way, provinces with similar spatial factors are a priori more likely to be clustered together. Posterior predictive inference in this model formalizes the desired prediction. Further, the identifiability of the proposed model is analyzed, and sufficient conditions to guarantee “generic” identifiability are provided. An$L_{1}$-penalized estimator is developed to assist variable selection and robust estimation when the number of explanatory covariates is large. An efficient expectation-minimization algorithm is presented for parameter estimation. Simulation studies and real-data examples are presented to investigate the empirical performance of the proposed method. Finally, it is worth noting that the proposed model has a wide range of practical applications, e.g., health management, environmental science, ecological studies, and so on.

Automated Driving Systems ( ADS ) have made great achievements in recent years thanks to the efforts from both academia and industry. A typical ADS is composed of multiple modules, including sensing, perception, planning, and control, which brings together the latest advances in different domains. Despite these achievements, safety assurance of ADS is of great significance, since unsafe behavior of ADS can bring catastrophic consequences. Testing has been recognized as an important system validation approach that aims to expose unsafe system behavior; however, in the context of ADS, it is extremely challenging to devise effective testing techniques, due to the high complexity and multidisciplinarity of the systems. There has been great much literature that focuses on the testing of ADS, and a number of surveys have also emerged to summarize the technical advances. Most of the surveys focus on the system-level testing performed within software simulators, and they thereby ignore the distinct features of different modules. In this article, we provide a comprehensive survey on the existing ADS testing literature, which takes into account both module-level and system-level testing. Specifically, we make the following contributions: (1) We survey the module-level testing techniques for ADS and highlight the technical differences affected by the features of different modules; (2) we also survey the system-level testing techniques, with focuses on the empirical studies that summarize the issues occurring in system development or deployment, the problems due to the collaborations between different modules, and the gap between ADS testing in simulators and the real world; and (3) we identify the challenges and opportunities in ADS testing, which pave the path to the future research in this field.

Class-imbalance is a prevalent and challenging problem in the field of fault detection. The undersampling ensemble framework is an effective method to deal with imbalance problems. However, designing a suitable sampling strategy to generate effective and divergent subsets is a major difficulty of this type of method. Hence, we propose a novel adaptive undersampling framework. It models the entire training process as a Markov decision process (MDP), thus, enabling dynamic decision-making for subsequent sampling strategies based on the current training performance of the ensemble framework. The sampler is optimized by the soft actor-critic reinforcement learning method. Considering the imbalance dataset's nature and the need for state definition, the clustering method is applied to the original training dataset. The state (training performance) and the action (sampling strategy) are determined according to the clustering results. The unique state definition and sampling decision mechanism are designed to ensure the convergence speed of MDP and improve the divergence of the subsets. We validate the effectiveness of the proposed framework on the real-world wind turbine blade cracking datasets and the high-speed train braking system dataset. The experimental results show that the classification performance and robustness of the proposed framework are significantly better than the 16 benchmark methods.

Fault diagnosis is an essential means to ensure the regular operation of mechanical systems. The existing data-driven algorithms are developed based on the assumption that the given label is entirely correct. However, mislabeling is common, which often occurs in industrial applications. These methods will overfit these mislabeled samples, resulting in inferior generalization. To this end, this article proposes a novel multistage true label distribution learning algorithm. Specifically, based on the training characteristics of data-driven algorithms on noisy datasets, a novel multistage adversarial loss function (MSA-Loss) is proposed. MSA-Loss can make the model construct the true label distribution from noisy datasets, prevent the model from overfitting the noisy samples, and finally keep the model with good generalization. The proposed method can be easily applied to any existing data-driven algorithm to improve its performance on noisy datasets. Our method is verified on high-speed aeronautical bearing and motor datasets, which prove that MSA-Loss has an excellent performance in noisy label scenarios. It can significantly improve the potential of existing diagnostic models in practical industrial applications.

Opportunistic maintenance (OM), which shows its superiority on complex multi-component systems by integrating the maintenance activities of multiple components to reduce the maintenance cost, has been widely studied over the past decade. To our knowledge, most of the existing OM works are developed based on fixed maintenance thresholds without fully utilizing the health state of the multi-component system. This article presents an OM optimization problem of multi-component systems with load sharing, solved by a modified proximal policy optimization approach based on deep reinforcement learning algorithm. The load sharing effect is reflected in the hazard rate function, which further changes the failure probability of the components. Meanwhile, the health states can be recovered by executing imperfect maintenance and corrective maintenance. The optimization problem is formulated as an infinite-horizon MDP with mixed discrete and continuous state and action space to maximize the total discounted reward, taking into account the system reliability and the maintenance cost. The difficulty caused by the mixed action space is solved by designing a parameterized action space structure and multi-task reinforcement learning framework. The effectiveness of the proposed algorithm is tested on a four-component system and a real-world scenario configured with the high-pressure feedwater heater system in the nuclear power plant. The results show that the performance of the algorithm is stable when facing large-scale problems. The algorithm proposed in this study also contributes to the imperfect maintenance optimization with state-of-the-art optimization techniques.

We propose a game attack–defense graph (GADG) approach that integrates the attack–defense graph and the game theory to model and analyze cyberattacks and defenses in the local metering system (LMS). Different from previous studies concentrating on static analyses of cybervulnerabilities, the GADG method considers correlations among these vulnerabilities. Besides, to avoid the uncertainty brought by unilateral analysis, we introduce the game theory for the interaction analysis between the attacker and the defender. The mixed-strategy Nash equilibrium that they finally reach can serve as the input for the inference of system states. We also propose two eliminating algorithms to reduce the complexity of solving mixed-strategy Nash equilibrium on large-scale LMS. In the case study, our GADG model has been conducted on a real LMS, and the results prove its efficiency. This research aims to assist the power company in optimizing the allocation of limited defensive resources, especially specific in attacking steps.Note to Practitioners—This article is motivated by the practical demand of protecting the cybersecurity of the local metering system (LMS), which is a critical subsystem of the smart grid and has been widely deployed to collect and transmit time-stamped information. Different from typical research works on local metering cybersecurity, which focuses on the analysis and identification of cybervulnerabilities, in this study, real experiments considering the interactive scenarios between the attacker and the defender are conducted on the LMS. Based on this, we developed the attack graph to clearly illustrate various cyberattack paths and their corresponding effects. We then applied the game theory model on the attack graph to obtain the optimal attack and defense strategy under each attack scenario. Based on this, the defender, i.e., power company, can optimally allocate limited defense resources among various attacking steps according to our model. The proposed model and method are demonstrated on one exemplar LMS used by our industrial partner China Southern Grid (CSG). Given the generalizability of the proposed model and method, they can be applied to other similar cyber–physical systems.

The papers in this special section focus on increasing interests in the development and implementation of advanced artificial intelligence (AI) and machine learning (ML) methods for tackling the reliability and system health prognostics challenges in various industrial applications.

Intelligent fault detection methods based on deep learning have been developed rapidly in recent years. However, most of these methods are based on supervised learning which requires a fully labeled training set. It is difficult to obtain massive labeled samples in real applications incredibly accurately labeled fault samples from an operating system. The lack of labels and label noise becomes a great challenge for fault detection. To tackle this problem, in this article, we propose a positive-unlabeled learning based hybrid network (PUHN). It only needs part of the normal operating samples to be labeled. All other samples (including the rest of the normal samples and all fault samples) are unlabeled, which greatly reduces the labeling cost. PUHN consists of three modules: a nonnegative risk positive-unlabeled (PU) network for training the classifier, a feature extraction module, and a clustering layer for improving data separability and estimating the class priors of PU learning. The three are optimized as a whole and the corresponding optimization strategy is designed. The monitoring data of 24 wind turbines are used to verify the effectiveness and robustness of the proposed method. The experimental results indicate that the proposed method is superior to the benchmark methods, and the performance is significantly better than the supervised learning method when there exists label noise.

With the lack of failure data, class imbalance has become a common challenge in the fault diagnosis of industrial systems. The oversampling methods can tackle the class-imbalanced problem by generating the minority samples to balance the training set. However, one of the main challenges of the existing oversampling methods is how to generate high-quality minority samples. Traditional oversampling methods regard all synthetic samples as minority ones to be added to the training set without filtering. The low-quality synthetic samples would distort the distribution of the dataset and worsen the classification performance. In this article, we propose a weakly supervised oversampling method that treats all synthetic samples as unlabeled samples and develops a graph semisupervised learning algorithm to select high-quality synthetic samples, adding into the final training set as minority samples. To improve the quality of synthetic samples, we propose a cost-sensitive neighborhood component analysis dimensionality reduction method to enhance domain information validity in high-dimensional datasets. Finally, combining a boosting-based ensemble framework, we propose a new imbalanced learning framework suitable for high dimensionality and highly imbalanced fault diagnosis in industrial systems. The experimental validation is performed on five real-world wind turbine blade cracking failure datasets and compared to 15 benchmark methods. The experimental results show that average performances and robustness of the proposed framework are significantly better than those of the benchmark methods.

In this paper, we consider a redundancy allocation problem for a series parallel system with uncertain component lifetimes that minimizes system costs while safeguarding system reliability over a given threshold level. We consider mixed redundancy strategies of cold standby and active redundancy with multiple types of components. We address lifetime uncertainty in the framework of distributionally robust optimization. In particular, we assume the probability distributions of the component lifetimes are not exactly known with only limited distributional information (e.g., mean, dispersion, and support) being available. We protect the worst-case system reliability constraint over all the possible component lifetime distributions that are consistent with the given distributional characteristics. The proposed modeling framework enjoys computationally attractive structures. The evaluation of the worst-case system reliability in our redundancy allocation problem can be transformed into a linear program, and the resulting overall redundancy allocation optimization problem can be cast as a mixed integer linear program that does not induce any additional integer variables (other than original allocation variables). In addition, the extreme joint distribution of component lifetimes can be efficiently recovered by solving a linear program. Our modeling framework can also be extended to incorporate the startup failures and common-cause failures for cold standbys and active parallels, respectively, to cater to more computationally complex settings. Finally, the computational experiments positively demonstrate the performance of the proposed approach in protecting system reliability.

Test-suite minimization is one key technique for optimizing the software testing process. Due to the need to balance multiple factors, multi-criteria test-suite minimization (MCTSM) becomes a popular research topic in the recent decade. The MCTSM problem is typically modeled as integer linear programming (ILP) problem and solved with weighted-sum single objective approach. However, there is no existing approach that can generate sound (i.e., being Pareto-optimal) and complete (i.e., covering the entire Pareto front) Pareto-optimal solution set, to the knowledge of the authors. In this work, we first prove that the ILP formulation can accurately model the MCTSM problem and then propose the multi-objective integer programming (MOIP) approaches to solve it. We apply our MOIP approaches on three specific MCTSM problems and compare the results with those of the cutting-edge methods, namely, NonlinearFormulation_LinearSolver (NF_LS) and two Multi-Objective Evolutionary Algorithms (MOEAs). The results show that our MOIP approaches can always find sound and complete solutions on five subject programs, using similar or significantly less time than NF_LS and two MOEAs do. The current experimental results are quite promising, and our approaches have the potential to be applied for other similar search-based software engineering problems.

The optimal feature selection problem in software product line is typically addressed by the approaches based on Indicator-based Evolutionary Algorithm (IBEA). In this study we first expose the mathematical nature of this problem --- multi-objective binary integer linear programming. Then, we implement/propose three mathematical programming approaches to solve this problem at different scales. For small-scale problems (roughly less than 100 features), we implement two established approaches to find all exact solutions. For medium-to-large problems (roughly, more than 100 features), we propose one efficient approach that can generate a representation of the entire Pareto front in linear time complexity. The empirical results show that our proposed method can find significantly more non-dominated solutions in similar or less execution time, in comparison with IBEA and its recent enhancement (i.e., IBED that combines IBEA and Differential Evolution).

A modeling and optimization framework for the maintenance of systems under epistemic uncertainty is presented in this paper. The component degradation processes, the condition-based preventive maintenance, and the corrective maintenance are described through piecewise-deterministic Markov processes in consideration of degradation dependence among degradation processes. Epistemic uncertainty associated with component degradation processes is treated by considering interval-valued parameters. This leads to the formulation of a multi-objective optimization problem whose objectives are the lower and upper bounds of the expected maintenance cost, and whose decision variables are the periods of inspections and the thresholds for preventive maintenance. A solution method to derive the optimal maintenance policy is proposed by combining finite-volume scheme for calculation, differential evolution, and nondominated sorting differential evolution for optimization. An industrial case study is presented to illustrate the proposed methodology.

Analogy-based software effort estimation is a method to estimate the project cost of an unseen project based on analogies against previous projects sharing selected features. The validity of the selected features depends on many factors, and one of most crucial factors is the effectiveness of the datapreprocessing techniques applied to the datasets of the previous projects. In this paper, we report the first controlled experiment that studies the class of three-stage data-preprocessing techniques with stages of missing data imputation, data normalization, and feature selection for analogy-based effort estimation. We conducted our investigation on the ISBSG data. The experimental results show that three-stage data-preprocessing techniques have significant impacts on the resultant effort estimation accuracy. The results also indicate that the combined use of Z-Score normalization, kNN imputation and mutual information based feature weighting can be an effective choice for analogy-based effort estimation.

Software quality prediction is an important yet difficult problem in software project development and management. Historical datasets can be used to build models for software quality prediction. However, the missing data significantly affects the prediction ability of models in knowledge discovery. Instead of ignoring missing observations, we investigate and improve incomplete-case k-nearest neighbor based imputation. K-nearest neighbor imputation is widely applied but has rarely been improved to have the most appropriate parameter settings for each imputation. This work conducts imputation on four well-known software quality datasets to discover the impact of the new imputation method we proposed. We compare it with mean imputation and other commonly used versions of k-nearest neighbor imputation. The empirical results show that the proposed dynamic incomplete-case nearest neighbor imputation performs better when the missingness is completely at random or non-ignorable, regardless of the percentage of missing values.

Components are often subject to multiple competing degradation processes. For multicomponent systems, the degradation dependence within one component or/and among components need to be considered. Physics-based models and multistate models are often used for component degradation processes, particularly when statistical data are limited. In this paper, we treat dependence between degradation processes within a piecewise-deterministic Markov process (PDMP) modeling framework. Epistemic (subjective) uncertainty can arise due to the incomplete or imprecise knowledge about the degradation processes and the governing parameters, to take this into account, we describe the parameters of the PDMP model as fuzzy numbers. Then, we extend the finite-volume method to quantify the (fuzzy) reliability of the system. The proposed method is tested on one subsystem of the residual heat removal system of a nuclear power plant, and a comparison is offered with a Monte Carlo simulation solution the results show that our method can be most efficient.

Many engineering systems can perform their intended tasks with various levels of performance, which are modeled as multi-state systems (MSS) for system availability and reliability assessment problems. Uncertainty is an unavoidable factor in MSS modeling, and it must be effectively handled. In this work, we extend the traditional universal generating function (UGF) approach for multi-state system (MSS) availability and reliability assessment to account for both aleatory and epistemic uncertainties. First, a theoretical extension, named hybrid UGF (HUGF), is made to introduce the use of random fuzzy variables (RFVs) in the approach. Second, the composition operator of HUGF is defined by considering simultaneously the probabilistic convolution and the fuzzy extension principle. Finally, an efficient algorithm is designed to extract probability boxes ($p$-boxes) from the system HUGF, which allow quantifying different levels of imprecision in system availability and reliability estimation. The HUGF approach is demonstrated with a numerical example, and applied to study a distributed generation system, with a comparison to the widely used Monte Carlo simulation method.

Open source software (OSS) approach has become increasingly prevalent for software development. As the widespread utilization of OSS, the reliability of OSS products becomes an important issue. By simulating the testing and debugging processes of software life cycle, the rate-based queueing simulation model has shown its feasibility for closed source software (CSS) reliability assessment. However, the debugging activities of OSS projects are different in many ways from those of CSS projects and thus the simulation approach needs to be calibrated for OSS projects. In this paper, we first characterize the debugging activities of OSS projects. Based on this, we propose a new rate-based queueing simulation framework for OSS reliability assessment including the model and the procedures. Then a decision model is developed to determine the optimal version-updating time with respect to two objectives: minimizing the time for version update, and maximizing OSS reliability. To illustrate the proposed framework, three real datasets from Apache and GNOME projects are used. The empirical results indicate that our framework is able to effectively approximate the real scenarios. Moreover, the influences of the core contributor staffing levels are analyzed and the optimal version-updating times are obtained.

Multistate physics modeling (MSPM) of degradation processes is an approach proposed for estimating the failure probability of components and systems. This approach integrates multistate modeling, which describes the degradation process through transitions among discrete states (e.g., initial, microcrack, rupture, etc.), and physics modeling by (physics) equations that describe the degradation process within the states. In reality, the degradation process is non-Markovian, its transition rates are time-dependent, and the degradation is possibly influenced by uncertain external factors such as temperature and stress. Under these conditions, it is in general difficult to derive the state probabilities analytically. In this paper, we overcome this difficulty by building a simulation model supported by a stochastic Petri net representing the multistate degradation process. The proposed modeling approach is applied to the problem of a nuclear component undergoing stress corrosion cracking. The results are compared with those derived from the state-space enrichment Markov chain approximation method applied in a previous work of literature.

Computing software and systems have become increasingly large and complex. As their dependability and autonomy are of great concern, self-healing is an ongoing challenge. This paper presents an innovative model and technology to realize the self-healing function under the real-time requirement. The proposed approach, different from existing technologies, is based on a new concept defined as consequence-oriented diagnosis and healing. Derived from the new concept, a prototype model for proactive self-healing actions is presented. Then, a hybrid diagnosis tool is proposed that takes advantages from the Multivariate Decision Diagram, Fuzzy Logic, and Neural Networks, achieving an efficient, effective, accurate, and intelligent result. The consequence-oriented diagnosis and self-healing function is also implemented. The experimental results exhibit that the innovative system is very effective and precise in predicting the consequence, and in preventing resulting software and system failures.