active library

reliability-estimation-in-series-systems-model-selection

Model selection for reliability estimation in series systems with Weibull components: when can engineers safely use simpler models?

Started 2023 HTML

Resources & Distribution

Package Registries

Model Selection for Reliability Estimation in Series Systems

This repository contains the research paper, simulation code, and supplementary materials for studying model selection in series systems with Weibull-distributed component lifetimes.

Key Findings

When can reliability engineers safely use a simpler model?

  • For well-designed systems (components with similar failure characteristics), a reduced homogeneous-shape Weibull model is statistically indistinguishable from the full heterogeneous model—even with 30,000 observations
  • This means practitioners can confidently use the simpler model, which:
    • Halves the parameter count from 2m to m+1
    • Renders the system itself Weibull-distributed
    • Reduces estimator variance without sacrificing accuracy
  • However, deviations in even a single component’s shape parameter quickly provide evidence against the reduced model

Practical Guidelines

Divergence LevelCV of Shape ParametersRecommendation
Low< 10%Use reduced model confidently
Moderate10-20%Depends on sample size
High> 25%Use full heterogeneous model

Repository Structure

.
├── paper/                  # LaTeX source and figures
   ├── paper.tex          # Main manuscript
   ├── refs.bib           # Bibliography
   └── image/             # Figures (PDF)
├── results/               # Simulation code and data
   ├── 5_system_scale3/   # Scale parameter sensitivity
   ├── 5_system_shape3/   # Shape parameter sensitivity
   ├── lrt/               # Likelihood ratio test simulations
      ├── divergence/    # Type I error and power analysis
      ├── vary_m/        # Effect of system complexity
      ├── vary_p/        # Effect of masking probability
      └── vary_q/        # Effect of censoring level
   └── ...
├── docs/                  # GitHub Pages site
└── CLAUDE.md              # Development guidance

Building the Paper

cd paper
pdflatex paper.tex
bibtex paper
pdflatex paper.tex
pdflatex paper.tex

Or using latexmk:

cd paper
latexmk -pdf paper.tex

Dependencies

R Packages

  • wei.series.md.c1.c2.c3 - Weibull series system with masked data
  • algebraic.mle - Maximum likelihood estimation utilities
  • tidyverse, ggplot2, parallel, boot

Python Packages

  • matplotlib, seaborn, pandas, numpy

LaTeX

Standard distribution with amsmath, amsthm, graphicx, natbib, hyperref

Citation

@article{towell2024modelselection,
  title={Model Selection for Reliability Estimation in Series Systems},
  author={Towell, Alex},
  year={2024},
  note={Preprint}
}

This paper builds on Towell (2023), which developed the likelihood model for masked and censored series system data.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Author

Alex Towell lex@metafunctor.com

Discussion