Parallel Simulation Experiments on a Personal Computer

1. Auflage 2027
ISBN: 978-1-041-39444-0
Verlag: Taylor & Francis Ltd

Buch, Englisch, 152 Seiten, Format (B × H): 156 mm x 234 mm

Reihe: Chapman & Hall/CRC Data Science Series

Parallel simulation is usually presented as a supercomputing topic, wrapped in clusters, schedulers, and specialized languages. Parallel Simulation Experiments on a Personal Computer argues the opposite: the multicore laptop or desktop you already own is enough to run serious simulation experiments, quickly and reproducibly, in both R and Python.

The book is built around sim_template, a compact, extensible program presented in matching R and Python versions, that distributes a simulation across worker processes while giving each worker its own high-quality random number stream. From that foundation the book works through complete, honest examples, with timing tables, missteps, and diagnostics included, showing not just how to parallelize a simulation but how to trust the numbers that come back.

Key Features:

- Reproducible parallel random number streams built on the PCG64 generator, via dqrng in R and NumPy's SeedSequence in Python

- Two full case studies: bootstrap interval estimation for a segmented regression change point in R, and a SIRD epidemic model with inter-process migration in Python

- Practical chapters on counting cores, memory pressure, benchmarking properly, load balancing, debugging parallel code, and reproducibility

- A look beyond the personal computer (GPUs, cloud instances, and clusters) and at working alongside AI coding assistants

- Optional appendices explaining how random number generators work, from congruential generators and RANDU to PCG64

Written for applied statisticians, data scientists, and graduate students who use simulation in their work, the book assumes only basic statistics and some familiarity with R or Python. No parallel computing background is required.

Zielgruppe

Academic and Professional Reference

Autoren/Hrsg.

Eubank, Randall L.

Ismay, Chester

Fachgebiete

• Mathematik | Informatik Mathematik Stochastik

Inhaltsverzeichnis

1. Introduction: Why Parallel Simulation?. 2. Random Numbers and Simulation. 3. Parallel Simulation in R. 4. Parallel Simulation in Python. 5. Case Study: Segmented Regression with Bootstrapping in R. 6. Case Study: Simulating an Epidemic with Python. 7. Practical Considerations. 8. Into the Beyond. 9. Conclusion.

Über Autor(innen)

Randall L. Eubank is Professor Emeritus of Statistics at Arizona State University. His research spans nonparametric regression, smoothing splines, functional data analysis, and statistical computing. He is the author of Nonparametric Regression and Spline Smoothing, A Kalman Filter Primer, and, with Ana Kupresanin, Statistical Computing in C++ and R. Over a career of more than four decades, at institutions including Southern Methodist University, Texas A&M University, and Arizona State University, he has taught statistical theory, computing, and methods courses at every level and supervised doctoral students. His professional service includes appointments as an associate editor of the Journal of the American Statistical Association, coordinating editor of the Journal of Statistical Planning and Inference, a decade-long associate editorship in statistical computing, and reviewer for National Science Foundation panels in probability and statistics. His long-standing interest in the computational machinery beneath statistical methods (how random numbers are generated, how algorithms behave in floating point, and what modern hardware makes possible) is the thread that runs through this book. He lives in Las Cruces, New Mexico.

Chester Ismay is a data science educator, consultant, and software developer. He is a co-author of Statistical Inference via Data Science: A ModernDive into R and the Tidyverse (CRC Press, second edition 2025, with Albert Y. Kim and Arturo Valdivia), a co-creator of the infer and fivethirtyeight R packages, and the creator of the moderndive Python package. He has taught statistics, mathematics, computer science, and data science at Ripon College, Pacific University, and Reed College, and has developed data science curriculum and courses in a range of roles since leaving academia. He now delivers corporate training in R, Python, SQL, and AI-assisted workflows for organizations including Portland State University, DataCamp, Data Society, and O'Reilly, and consults with organizations on improving their data and AI workflows, with a focus on building critical thinking skills and greater efficiency. He holds a Ph.D. in Statistics from Arizona State University, where his dissertation was directed by his co-author. His work centers on making rigorous statistical computing approachable: clear templates, reproducible workflows, and tooling that meets practitioners where they are. He lives in Tempe, Arizona.