The stress-strength (SS) reliability metric is essential to reliability engineering and other scientific disciplines. It aids in forecasting failures of systems, materials, and structures. Using adaptive progressive hybrid censored data, this study evaluates the reliability of a system with two independent random variables, X and Y, which follow an inverse Pareto distribution (IPD). The random variables X and Y represent the strength and stress of the system. This study primarily aims to compare various estimates of SS reliability utilising both frequentist and Bayesian methodologies, specifically maximum likelihood (ML), maximum product spacing (MPS), and bootstrap (boot-p boot-t) methods for frequentist approaches, alongside MCMC and T–K approximation methods for Bayesian approaches. Bayesian computation presumes the gamma distribution for prior information and uses the squared error loss function for estimations. We perform extensive Monte Carlo simulations to determine the most accurate feasibility of the estimation techniques. The simulated results indicated that the Bayesian estimation approach using MCMC techniques outperformed all other estimate methods. In addition, the adopted methodologies have been fully applied with two illustrative examples.
Uncertainties and delays are inherent features of many natural phenomena, especially pharmacokinetics, since drug absorption and degradation may change from person to person and with time. These features naturally give rise to mathematical models involving uncertainty, memory and delayed responses, which can be effectively described using fuzzy mixed Volterra–Fredholm delay integro-differential equations. In this study, a variational iteration technique for finding the solutions to fuzzy mixed Volterra–Fredholm delay integro-differential equations is established, which can effectively capture both uncertainties and memory effects. By employing the α -cut methodology, the fuzzy equation is transformed into interval-valued equations and a powerful algorithm is developed by avoiding discretization. One important advance is the stepwise contraction analysis in the standard supremum norm, in which the influence of the delay effect is explicitly incorporated. This approach avoids the conventional global assumption LT<1 by establishing the contraction condition locally on successive subintervals. The proposed scheme ensures solution existence and uniqueness under the stated assumptions, while the VIM iterations converge systematically. Using the stopping criterion ‖ W_n+1-W_n‖ _∞≤ 10^-4 , the prescribed tolerance is reached within six iterations for the linear model and seven iterations for the nonlinear Michaelis–Menten model. To demonstrate its practical applicability, the proposed framework is applied to two pharmacokinetic models: a linear delayed model and a nonlinear Michaelis–Menten model.
File fragment type identification is a critical research topic in digital forensics. However, current methods still face numerous challenges, including low accuracy, reliance on manually designed features, and limited coverage of file types. To address these limitations, inspired by binocular vision principles, we propose a Parallel Cross Convolutional Neural Network (PCCNN) architecture. This architecture overcomes the limitations of single-structure design and insufficient feature extraction in existing models. The PCCNN can automatically learn features from raw bytes. It extracts features through parallel left and right branch networks, conducts multiple feature concatenation and convolution operations, and fuses feature information from different branches, thereby achieving more comprehensive feature extraction and improving model performance. On the FFT-75 dataset, our model outperforms FiFTy. Compared with the FiFTy model, the average accuracy across six scenarios is improved by 15
