Tianli Xiao

ORCID: 0009-0006-4003-8099
Publications
Citations
Views
---
Saved
---
About
Contact & Profiles
Research Areas
  • Probabilistic and Robust Engineering Design
  • Advanced Multi-Objective Optimization Algorithms
  • Optimal Experimental Design Methods
  • Fatigue and fracture mechanics
  • Non-Destructive Testing Techniques
  • Metaheuristic Optimization Algorithms Research
  • Structural Health Monitoring Techniques

Nanjing University of Science and Technology
2022-2023

Abstract Surrogate models have been proven to be powerful tools alleviate the computational burden of structural reliability analysis. An appropriate surrogate model can guarantee prediction accuracy with limited samples. However, traditional single modeling technique ignores model‐form uncertainty due insufficient knowledge physical system, leading unreliable results or time‐consuming computation. To overcome aforementioned deficiencies, an active learning ensemble under framework Bayesian...

10.1002/qre.3152 article EN Quality and Reliability Engineering International 2022-06-21

The uncertainty of design variables is usually ignored in deterministic design, which may lead to unreliable process parameters for actual production and unable provide satisfactory products. To ensure the higher reliability a product, novel robust method based on coordinated mechanism surrogate models proposed. In implementation process, Kriging model polynomial response surface are adjusted achieve optimization, ensuring that solution feasible within interval. A gear reducer utilized...

10.1109/icccr56747.2023.10194011 article EN 2023-03-24

Expensive multi-objective black-box optimization problems often only allow limited functional evaluations because of time and cost considerations. However, traditional evolutionary algorithms usually require many sample evaluations, making it challenging to solve the above expensive problems. Multi-objective Bayesian can efficiently by using a sequential search define infill criteria. Still, conventional serial computing ways cannot fully utilize abundant computer resources. To address...

10.1109/ccci58712.2023.10290829 article EN 2023-10-18

In statistical inference, experiment data are usually assumed to be normally distributed and uncontaminated. However, these assumptions may not hold in practice. this paper, we propose a robustified analogue 𝑧𝑧 score of the conventional 𝑡𝑡 which is constructed with two robust estimator Huber’s M-estimator (Huber) Rousseeuw Croux (RC). The properties estimators provided, asymptotic normality property proposed also proven theoretically. empirical power receiver operating characteristic curve...

10.2139/ssrn.4184980 article EN SSRN Electronic Journal 2022-01-01
Coming Soon ...