Article
Computer Science, Information Systems
Kai Cheng, Zhenzhou Lu
Summary: This paper introduces two new support vector regression (SVR) models under the Bayesian inference framework, which can provide point-wise probabilistic prediction and determine optimal hyperparameters. Numerical results show that these models are very promising for constructing accurate regression models for problems with diverse characteristics.
INFORMATION SCIENCES
(2021)
Article
Engineering, Multidisciplinary
Jinsheng Wang, Chenfeng Li, Guoji Xu, Yongle Li, Ahsan Kareem
Summary: An adaptive algorithm based on the Bayesian SVR model (ABSVR) is proposed in this study, which combines new learning functions, distance constraint terms, and adaptive sampling region schemes to improve efficiency and reliability.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2021)
Article
Engineering, Industrial
Haoyuan Shen, Yizhong Ma, Chenglong Lin, Jian Zhou, Lijun Liu
Summary: This paper proposes a hierarchical Bayesian support vector regression (HBSVR) model for dynamic high-dimensional reliability modeling, which combines the step-size adaptive accelerated Markov Chain Monte Carlo (SAA-MCMC) method with Sequential Minimal Optimization (SMO) for parameter calibration and dynamic update. The HBSVR model is further improved by applying an active learning algorithm to continuously improve model accuracy.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Engineering, Civil
Chao Dang, Marcos A. Valdebenito, Matthias G. R. Faes, Jingwen Song, Pengfei Wei, Michael Beer
Summary: This paper proposes a new method called 'Bayesian active learning line sampling' (BAL-LS), which derives the exact posterior variance of the failure probability to measure the epistemic uncertainty more accurately. In addition, the method proposes two essential components (learning function and stopping criterion) to facilitate Bayesian active learning based on the uncertainty representation of the failure probability. Compared with PBAL-LS, BAL-LS also has the advantage of automatically updating the important direction. Four numerical examples demonstrate the efficiency and accuracy of the proposed method in evaluating extremely small failure probabilities.
Article
Computer Science, Artificial Intelligence
Firat Ozdemir, Zixuan Peng, Philipp Fuernstahl, Christine Tanner, Orcun Goksel
Summary: This paper proposes an active learning framework that optimally utilizes expert clinician time in medical image analysis, generating improved segmentation performance. By combining representativeness with uncertainty, the method iteratively estimates ideal samples to be annotated from a given dataset.
KNOWLEDGE-BASED SYSTEMS
(2021)
Article
Engineering, Multidisciplinary
Tianli Xiao, Chanseok Park, Linhan Ouyang, Yizhong Ma
Summary: This paper proposes an active learning ensemble surrogate model under the framework of Bayesian inference for structural reliability analysis, aiming to alleviate the computational burden and ensure prediction accuracy with limited samples. By developing a learning function and an adaptive method, computational efficiency is greatly enhanced, and the effectiveness and robustness of the algorithm are verified in multiple application cases.
QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL
(2022)
Review
Engineering, Industrial
Atin Roy, Subrata Chakraborty
Summary: Support vector machine (SVM) is a powerful machine learning technique widely used in structural reliability analysis (SRA). This article provides a comprehensive review of various SVM approaches in SRA applications, including classification and regression algorithms. The article also discusses advanced variants of SVM and hyperparameter tuning algorithms. The review highlights the excellent capability of SVM in handling high-dimensional problems with relatively fewer training data in SRA applications.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2023)
Article
Computer Science, Information Systems
Zichen Zhang, Shifei Ding, Yuting Sun
Summary: This paper introduces a new method called multiple birth support vector regression (MBSVR), which constructs the regressor from multiple hyperplanes obtained by solving small quadratic programming problems, aiming for faster computation and better fitting precision.
INFORMATION SCIENCES
(2021)
Article
Engineering, Industrial
Nick Pepper, Luis Crespo, Francesco Montomoli
Summary: This work demonstrates how to approximate the failure probability of an expensive computational model with reliability requirements using Support Vector Machines. An algorithm is proposed to select informative parameter points to improve the approximation accuracy iteratively. Additionally, a method is provided to quantify the uncertainty in the Limit State Function and estimate an upper bound to the failure probability using geometrical arguments.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2022)
Article
Engineering, Aerospace
Jiangfeng Fu, Fangqi Hong, Pengfei Wei, Zongyi Guo, Yuannan Xu, Weikai Gao
Summary: Due to limited information at the early design stage, evaluating the reliability of aerospace structures accurately is challenging. Imprecise probability models are widely developed and accepted for separating uncertainties, but propagating these models through expensive simulators is difficult. To address this, a Bayesian active learning method is proposed for efficiently learning failure probability and response variance. This method is effective and provides accurate results.
AEROSPACE SCIENCE AND TECHNOLOGY
(2023)
Article
Computer Science, Artificial Intelligence
Ying Liu, Yifei Wang, Long Chen, Jun Zhao, Wei Wang, Quanli Liu
Summary: An incremental Bayesian framework broad learning system is proposed to efficiently reduce the scale of matrix operations and achieve better outcomes in experiments compared to traditional BLS and other comparative algorithms through incremental learning.
ARTIFICIAL INTELLIGENCE REVIEW
(2021)
Article
Engineering, Mechanical
Atin Roy, Subrata Chakraborty, Sondipon Adhikari
Summary: Adaptive sampling near a limit state is crucial for reliability analysis of structures using metamodels. Active learning based on posterior mean and standard deviation provided by a chosen metamodel is widely used for such adaptive sampling. In this study, active learning-enhanced adaptive sampling-based sparse Bayesian regression is explored for reliability analysis.
JOURNAL OF ENGINEERING MECHANICS
(2023)
Article
Construction & Building Technology
I-Tung Yang, Handy Prayogo
Summary: Reliability-based design optimization considers uncertainties in the designing process of resilient buildings and structures. A new framework, called SOS-ASVM, is proposed to classify structural designs and achieve better accuracy. The framework integrates SOS and ASVM, and has been validated in three practical engineering cases.
Article
Computer Science, Interdisciplinary Applications
Tong Zhou, Yongbo Peng
Summary: In this paper, an active-learning reliability method called ASVR-PDEM is proposed to reduce the computational cost in structural reliability analysis. The method combines the support vector regression (SVR) and the probability density evolution method (PDEM). The empirical probability distribution of the SVR is proposed based on the leave-one-out cross-validation (LOOCV) strategy, and two learning functions are proposed based on the PDEM-oriented expected improvement function. The results show that the ASVR-PDEM outperforms conventional reliability methods in terms of computational accuracy, efficiency, and computational time, especially for time-consuming dynamic reliability problems.
COMPUTERS & STRUCTURES
(2023)
Article
Engineering, Electrical & Electronic
Sheng-Tong Zhou, Jian Jiang, Jian-Min Zhou, Pei-Han Chen, Qian Xiao
Summary: The accuracy of the BSVR metamodel is insufficient for analyzing uncertainty in complex systems using the traditional one-shot sampling method. To address this issue, this study presents an error-pursuing adaptive uncertainty analysis method based on the BSVR metamodel, incorporating a new adaptive sampling scheme. The new sampling scheme utilizes an adjusted mean square error (AMSE) function to guide the adaptive sampling process, estimating the prediction error of the metamodel. The proposed method was validated using benchmark analytical functions and demonstrated its effectiveness in a realistic application of an overhung rotor system.
Article
Engineering, Industrial
Mateusz Oszczypala, Jakub Konwerski, Jaroslaw Ziolkowski, Jerzy Malachowski
Summary: This article discusses the issues related to the redundancy of k-out-of-n structures and proposes a probabilistic and simulation-based optimization method. The method was applied to real transport systems, demonstrating its effectiveness in reducing costs and improving system availability and performance.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Wencheng Huang, Haoran Li, Yanhui Yin, Zhi Zhang, Anhao Xie, Yin Zhang, Guo Cheng
Summary: Inspired by the theory of degree entropy, this study proposes a new node identification approach called Adjacency Information Entropy (AIE) to identify the importance of nodes in urban rail transit networks (URTN). Through numerical and real-world case studies, it is found that AIE can effectively identify important nodes and facilitate connections among non-adjacent nodes.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Hongyan Dui, Yaohui Lu, Liwei Chen
Summary: This paper discusses the four phases of the system life cycle and the different costs associated with each phase. It proposes an improvement importance method to optimize system reliability and analyzes the process of failure risk under limited resources.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Xian Zhao, Chen Wang, Siqi Wang
Summary: This paper proposes a new rebalancing strategy for balanced systems by switching standby components. Different switching rules are provided based on different balance conditions. The system reliability is derived using the finite Markov chain imbedding approach, and numerical examples and sensitivity analysis are presented for validation.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Fengyuan Jiang, Sheng Dong
Summary: Corrosion defects are the primary causes of pipeline burst failures. The traditional methodologies ignore the effects of random morphologies on failure behaviors, leading to deviations in remaining strength estimation and reliability analysis. To address this issue, an integrated methodology combining random field, non-linear finite element analysis, and Monte-Carlo Simulation was developed to describe the failure behaviors of pipelines with random defects.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Guoqing Cheng, Jiayi Shen, Fang Wang, Ling Li, Nan Yang
Summary: This paper investigates the optimal joint inspection and mission abort policies for a multi-component system with failure interaction. The proportional hazards model is used to characterize the effect of one component's deterioration on other components' hazard rates. The optimal policy is studied to minimize the expected total cost, and some structural properties of the optimal policy are obtained.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Hongyan Dui, Yaohui Lu, Shaomin Wu
Summary: A new resilience model is proposed in this paper for systems under competing risks, and related indices are introduced for evaluating the system's resilience. The model takes into account the degradation process, external shocks, and maintenance interactions of the system, and its effectiveness is demonstrated through a case study.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Yang Li, Jun Xu
Summary: This paper proposes a translation model based on neural network for simulating non-Gaussian stochastic processes. By converting the target non-Gaussian power spectrum to the underlying Gaussian power spectrum, non-Gaussian samples can be generated.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Yanyan Liu, Keping Li, Dongyang Yan
Summary: This paper proposes a new random walk method, CBDRWR, to analyze the potential risk of railway accidents. By combining accident causation network, we assign different restart probabilities to each node and improve the transition probabilities. In the case study, the proposed method effectively quantifies the potential risk and identifies key risk sources.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Nan Hai, Daqing Gong, Zixuan Dai
Summary: The current risk management of utility tunnel operation and maintenance is of low quality and efficiency. This study proposes a theoretical model and platform that offer effective decision support and improve the safety of utility tunnel operation and maintenance.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Tomoaki Nishino, Takuya Miyashita, Nobuhito Mori
Summary: A novel modeling methodology is proposed to simulate cascading disasters triggered by tsunamis considering uncertainties. The methodology focuses on tsunami-triggered oil spills and subsequent fires and quantitatively measures the fire hazard. It can help assess and improve risk reduction plans.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Mingjiang Xie, Yifei Wang, Jianli Zhao, Xianjun Pei, Tairui Zhang
Summary: This study investigates the effect of rockfall impact on the health management of pipelines with fatigue cracks and proposes a crack propagation prediction algorithm based on rockfall impact. Dynamic SIF values are obtained through finite element modeling and a method combining multilayer perceptron with Paris' law is used for accurate crack growth prediction. The method is valuable for decision making in pipeline reliability assessment and integrity management.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Saeed Jamalzadeh, Lily Mettenbrink, Kash Barker, Andres D. Gonzalez, Sridhar Radhakrishnan, Jonas Johansson, Elena Bessarabova
Summary: This study proposes an integrated epidemiological-optimization model to quantify the impacts of weaponized disinformation on transportation infrastructure and supply chains. Results show that disinformation targeted at transportation infrastructure can have wide-ranging impacts across different commodities.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Jiaxi Wang
Summary: This paper investigates the depot maintenance packet assignment and crew scheduling problem for high-speed trains. A mixed integer linear programming model is proposed, and computational experiments show the effectiveness and efficiency of the improved model compared to the baseline one.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)
Article
Engineering, Industrial
Yuxuan Tian, Xiaoshu Guan, Huabin Sun, Yuequan Bao
Summary: This paper proposes a DFMs searching algorithm based on the graph neural network (GNN) to improve computational efficiency and adaptively identify DFMs. The algorithm terminates prematurely when unable to identify new DFMs.
RELIABILITY ENGINEERING & SYSTEM SAFETY
(2024)