https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/209 2026-07-04T11:28:49Z https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/212275 Title: Unified stochastic modeling and reliability assessment for coupled degradation mechanisms Authors: Tian, Runcao; Bae, Suk Joo; Chen, Zhongshu; Liu, Yu Abstract: Engineered systems often experience coupling effects with multiple degradation mechanisms during their operation. A well-tuned modeling process of such complex degradation mechanisms is crucial for accurate reliability assessment of engineered systems. This study puts forth a unified stochastic process model for (accelerated) degradation data with coupled mechanisms in a form of weighted mixture. The weighted hybrid degradation process is based mainly on the Tweedie exponential dispersion process (TEDP) as a unified model of traditional stochastic processes for a unique degradation mechanism via a continuously adjustable shape parameter and nonlinear time transformation. By assigning proper weights to quantify the contributions of coupled degradation effects, the weighted mixture model permits flexible modeling of complex degradation mechanisms. Under the proposed degradation modeling framework, we propose a new accelerated degradation test (ADT) model to extrapolate lifetime distribution at normal use condition through the relationship between stress and degradation rate. To derive maximum likelihood estimates (MLEs) of the model parameters, we newly design the expectation-maximization (EM) algorithm and compare the performance of widely adopted numerical optimizers in terms of convergence rate and computational efficiency. A variety of simulation studies and analyses of two real-world cases validate the effectiveness of the proposed degradation modeling and reliability assessment framework. 2026-08-01T00:00:00Z https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/214012 Title: Two-phase optimal and heuristic algorithms for flow shop scheduling with reworks under overlapped queue time limits Authors: Kim, Hyeon-Il; Lee, Dong-Ho Abstract: This study addresses flow shop scheduling in which each job is reworked when one of its overlapped queue time limits is violated. The problem is to determine the start times of jobs at each stage and rework setups/operations if occur. After decomposing the problem into the job process route selection and the resulting flow shop or reentrant flow shop scheduling sub-problems, a two-phase optimal branch and bound (TP-B&B) algorithm is proposed for the three-stage makespan problem. The algorithm consists of generating the job process routes using a two-level tree, and finding the resulting optimal schedules using a B&B algorithm while reducing the search space by a dominance property and the best upper bound obtained by solving the resulting scheduling problems. Moreover, for the multi-stage makespan problem, a basic two-phase variable neighborhood search (TP-VNS) algorithm is proposed that solves the two sub-problems iteratively, where each sub-problem is solved using a shaking and local search improvement method with different neighborhood structures. Then, it is extended to a general TP-VNS algorithm with a variable neighborhood descent method. Computational results show that the TP-B&B algorithm gave 21.5% more optimal solutions than the Gurobi while requiring much less computation times for the small sized test instances. Also, the TP-VNS algorithms outperform the existing one significantly. Specifically, the basic (general) TP-VNS algorithm gave 0.7% (1.2%) improvement in the overall optimality gap for small sized test instances, and 1.2% (2.9%) improvement in the overall relative performance ratio for medium-to-large sized test instances. 2026-07-01T00:00:00Z https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/212385 Title: A spatio-temporal Gaussian process with change-points for image-based degradation data Authors: Lim, Munwon; Bae, Suk Joo Abstract: Condition monitoring and fault diagnosis using big-data analytic and artificial intelligence (AI) have been an essential tool for reliable operation and timely maintenance of machines, facilitating condition-based maintenance (CBM). To continuously monitor the health of machining tools, we propose a spatio-temporal Gaussian process with change-points (CP-STGP) for modeling image-based degradation patterns. Introducing the concept of change-points, we aim to detect degradation transitions in space-time domain to determine the optimal replacement time for machining tools. The proposed model adopts spectral representation through partial derivatives for complex correlation structure of covariance function in space-time domain. We also introduce the Kalman filter algorithm after transforming original image data using the fast Fourier transform to attempt computational efficiency of re-parameterization. By sequentially applying the likelihood-ratio test (LRT) for multivariate normal models, we derive maximum likelihood estimates (MLEs) of the parameters of the CP-STGP model by determining the number of change-points a priori. Inference on the model parameters is then derived, based on the asymptotic distribution of the LRT statistic. The analysis of a cylinder system in an automobile and simulation results show that the CP-STGP model effectively captures varying image patterns over time by separately modeling them before and after change-points. The proposed modeling approach is expected to help maintenance engineers automatically determine the best replacement time for machining tools as an alternative to manual inspection in practice. 2026-06-01T00:00:00Z https://scholarworks.bwise.kr/hanyang/handle/2021.sw.hanyang/210637 Title: Learning standardized noise for risk-neutral option pricing via Generative Adversarial Networks Authors: Choi, Young Hoon; Ryu, Dongwon; Byun, Jun Young; Na, Yosep; Song, Jae Wook Abstract: This paper proposes a Generative Adversarial Network (GAN)–based framework for risk-neutral option pricing that learns the empirical distribution and temporal structure of log-return noise. By extracting and modeling stochastic noise from historical returns, the framework generates risk-neutral price paths for option valuation and delta prediction. We evaluate three state-of-the-art GAN architectures, including TimeGAN, QuantGAN, and SigCWGAN, against Monte Carlo simulation, the Black–Scholes–Merton, and Heston models across market regimes, maturities, moneyness levels, and option types. Empirical results show that QuantGAN and SigCWGAN accurately replicate key distributional and autocorrelation properties of return noise and consistently outperform benchmark models in option pricing, particularly in stable market environments and around at-the-money regions where pricing accuracy is most critical. Across a broad range of market conditions, these models deliver lower pricing errors and higher statistical confidence measures than traditional benchmarks. While pricing performance deteriorates during periods of abrupt volatility shifts, GAN-based models remain competitive overall. In contrast, improvements in delta prediction are limited, especially near mid-delta regions where payoff curvature is steepest. Overall, the findings demonstrate that learning stochastic noise offers an effective and flexible data-driven alternative for risk-neutral option pricing, while reliable sensitivity estimation requires models that jointly capture distributional features and local dynamic responses of the underlying asset. 2026-03-01T00:00:00Z