A5035165512- Advanced Statistical Methods and Models
- Advanced Statistical Process Monitoring
- Software Reliability and Analysis Research
- Low-power high-performance VLSI design
- Mathematics and Applications
- Optimal Experimental Design Methods
- Reliability and Maintenance Optimization
- Scientific Measurement and Uncertainty Evaluation
- Probabilistic and Robust Engineering Design
- Software Engineering Research
- VLSI and Analog Circuit Testing
- Fault Detection and Control Systems
- Polynomial and algebraic computation
- Advanced Combinatorial Mathematics
- Advanced Mathematical Identities
- Machine Fault Diagnosis Techniques
- Advanced Topics in Algebra
- Software Testing and Debugging Techniques
- Matrix Theory and Algorithms
- Mathematical functions and polynomials
- Bayesian Methods and Mixture Models
- graph theory and CDMA systems
- Algebraic structures and combinatorial models
- Sensory Analysis and Statistical Methods
- Manufacturing Process and Optimization
Eindhoven University of Technology
2015-2024
University of Maryland, College Park
2018
Centraal Bureau voor de Statistiek
1994-2007
Tilburg University
2001
Université de Bordeaux
1994
University of Groningen
1991-1994
Variability is an important aspect of SRAM cell design. Failure probabilities P <inf xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">fail</inf> ≤10 <sup xmlns:xlink="http://www.w3.org/1999/xlink">−10</sup> have to be estimated through statistical simulations. Accurate techniques such as Importance Sampling Monte Carlo simulations are essential accurately and efficiently estimate low failure probabilities. This paper shows that a simple form...
In modern industrial systems, sensor data reflecting the system health state are commonly used for remaining useful lifetime (RUL) prediction, which increasingly processed by deep learning based approaches recently. But these models do not automatically provide uncertainty information RUL hence this paper is motivated to introduce a novel approach that allows control trade-off between prediction performance and knowledge about of prediction. The key aspect our use long short-term memory...
There is a variation in the characteristics of an LED partly due to manufacturing process and sourcing materials from differing vendors. From thermal perspective these variations result such as R <sub xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">thJC</sub> (the resistance between chip junction case) Z xmlns:xlink="http://www.w3.org/1999/xlink">th</sub> impedance curve). This study seeks quantify measured number part samples provide insight into...
We present a new, natural way to construct nonparametric multivariate tolerance regions. Unlike the classical intervals, where endpoints are determined by beforehand chosen order statistics, we take shortest interval that contains certain number of observations. extend this idea higher dimensions replacing class intervals general indexing sets, which specializes classes ellipsoids, hyperrectangles or convex sets.The asymptotic behavior our regions is derived using empirical process theory,...
A time-domain methodology for statistical simulation of nonlinear dynamic integrated circuits with arbitrary excitations is presented. The behavior the described as a set stochastic differential equations rather than estimated by population realizations and Gaussian closure approximations are introduced to obtain closed form moment equations. Statistical specific shows that proposed numerical methods offer accurate efficient solution differentials variability noise analysis circuits.
The full potential of predictive maintenance has not yet been utilised. Current solutions focus on individual steps the cycle and only work for very specific settings. overarching challenge is to leverage these building blocks obtain a framework that supports optimal asset management. PrimaVera project identified four obstacles tackle in order utilise at its potential: lack orchestration automation workflow, inaccurate or incomplete data role human organisational factors data-driven decision...
It is known that quantitative measures for the reliability of software systems can be derived from models, and, as such, support product development process. Over past four decades, research activities in this area have been performed. As a result, many models proposed. was shown that, once these reach certain level convergence, it enable developer to release and stop testing accordingly. Criteria determine optimal time include number remaining errors, failure rate, requirements, or total...
Abstract We present a case study in monitoring high‐volume production process with high yield. Testing the products is difficult and only possible destructive way defect/no‐defect result. review several attribute charting procedures for high‐yield processes. It turns out that number of conforming items between occurrence non‐conforming appropriate. discuss implementation aspects CCC r charts our study. The based on new formula standard deviation inspected before signal. Copyright © 2005 John...
We show how to use computer algebra for computing exact distributions of nonparametric test statistics. give several examples statistics with explicit probability-generating functions that can be handled this way. In particular, we a new table critical values the Jonckheere–Terpstra extends previous tables.
This report describes work performed during SWI 2023 at the University of Groningen in relation with Problem 1 posed by company ASMPT. They have detailed simulation software a machine and they compare results this physical experimental results. There is significant difference between simulated measured data, it goal to study how estimate parameters model using ex-perimentally frequency response. First, two toy models are studied understand challenges pa- rameter estimation domain. Later,...
Abstract Statistical process monitoring of high‐purity manufacturing processes becomes challenging if the defect rate depends on fluctuations a set covariates (e.g., inspected weight, volume, temperature). This paper applies generalized linear model framework to statistical control for detecting contextual anomalies in processes. Different types predictive residuals (i.e., Pearson, deviance, and quantile) recursive are considered, performance these schemes is compared via simulation study.
Abstract An increasing number of applications in the chemical industry involve measuring nonconforming items, particularly high‐purity processes or high‐yield processes. Dedicated monitoring tools such as time‐between‐events (TBE) control charts have been developed for both discrete time (CCC‐charts) and continuous ( t r ‐charts) detecting any shifts process defect rate. However, most common performance metrics used literature are not always appropriate may suffice to describe efficiency a...
For robust design of SRAM memories, it is not sufficient to guarantee good statistical margins on the cell parameters. The sense amplifier needs input signal before can reliably data, while requires time develop that signal. This paper presents a new method allows optimization access an memory, guaranteeing yield target set by designer. Using this method, high performance advanced CMOS has been improved 6%, simultaneously reducing size.
Abstract Generalized likelihood ratio (GLR) control charts are useful for tailor‐made monitoring strategies, but they less developed discrete processes. In this paper, the GLR chart framework applied to aggregate cumulative quantities data is extended. Inspired by technical note on from Lee and Woodall (2018), unnecessary artificial bounds in geometric proposed literature removed parameter restriction errors, common designs, corrected. Finally, Gamma continuous‐time time‐between‐event that...