Who uses PyDOE?¶
PyDOE is a widely used Python library for design of experiments (DOE). Its users range from individual researchers to global organizations, spanning academia, industry, and open-source projects. It supports applications in simulation-based optimization, engineering workflows, scientific research, and data-driven model development.
Selected Testimonials
Below are some small selection of the many institutes and companies that use PyDOE.
GlennOPT is a Python optimization tool developed by NASA for multi-objective optimization of engineering simulations (e.g., CFD problems). It uses evolutionary strategies (like Differential Evolution and NSGA methods) to explore design spaces, handle simulation failures, and restart long-running jobs robustly. It’s designed to track evaluations and integrate with external simulation workflows. It uses PyDOE to generate design-of-experiments samples for multi-objective simulation-based optimization.
Source: github.com/nasa/GlennOPT
IBM's SimulAI is a Python toolkit for physics-informed machine learning, combining models like Physics-Informed Neural Networks (PINNs), DeepONets, autoencoders, etc., for scientific computing and reduced-order modeling. It’s intended to unify state-of-the-art ML techniques for physical systems.
Source: github.com/ibm/simulai
Holodeck is an astrophysics software package developed by NANOGrav for massive black-hole binary population synthesis and for generating synthetic populations of supermassive black hole binaries for gravitational-wave background studies and related analyses.
Source: github.com/nanograv/holodeck
ARMI (Automated Reactor Modeling Infrastructure) is Terrapower’s Python framework for simulation and analysis of nuclear reactors’ core physics, including fuel performance and neutronics modeling. It builds complex reactor modeling pipelines.
Source: github.com/terrapower/armi
OpenMDAO is an open-source, multidisciplinary optimization framework in Python designed for systems engineering and analysis of complex coupled systems developed by NASA Glenn Research Center. It allows linking models, setting up design variables, and running optimizations with gradient and parallel support.
Source: github.com/OpenMDAO/OpenMDAO
Los Alamos National Laboratory' Bohydra is a workflow and data orchestration framework that supports design-of-experiments, optimization, UQ, and HPC workflows, particularly where multiple simulation codes must be coordinated. It’s focused on ensemble studies and model-based engineering.
Source: github.com/lanl/bohydra
Lawrence Livermore National Laboratory uses DOE in Zero-RK and Merlin-Spellbook to generate design-of-experiments samples for simulation-based optimization workflows. Zero-RK is a numerical ODE integrator / solver framework optimized for HPC and accurate time integration of ODE systems. It provides high-performance integrators for stiff and non-stiff problems, often used in scientific computing. Merlin-spellbook is a small utility package (often used alongside Merlin workflows) providing common building blocks for configuring workflows and tasks.
Sources:
github.com/llnl/zero-rk
github.com/llnl/merlin-spellbook
pvOps is a Python library for photovoltaic (PV) systems data analysis, particularly for field-collected operational data, text logs, time series, IV curves, etc. It includes processing and fusion of diverse PV datasets. Sandia National Laboratories uses DOE in pvOps create samples for black box optimization.
Source: github.com/sandialabs/pvOps
GEMSEO (Generic Engine for Multi-disciplinary Scenarios, Exploration and Optimization) is a general engine for multidisciplinary scenarios, exploration, and optimization — a suite for optimization, UQ, MDO, and surrogate modeling tools. It’s modular and acts as an engine for exploring engineering systems under uncertainty.
Source: github.com/gemseo/gemseo
The Multiobjective Optimization Group uses DOE in pyRVEA to produce DOE samples for multi-objective optimization and surrogate model workflows. It implements the Reference Vector Guided Evolutionary Algorithm for global search in design spaces.
Gators is a package developed by Paypal to handle model building with big data and fast real-time pre-processing, even for a large number of QPS, using only Python. It uses DOE to generate structured design samples for model testing and performance tuning. It handles large-scale model building and real-time preprocessing on high-QPS data pipelines, providing a systematic way to explore input spaces efficiently before training or benchmarking.
Source: github.com/paypal/gators
SURGroup uses DOE in UQpy (“Uncertainty Quantification with Python”) to create design-of-experiments samples for uncertainty quantification and optimization workflows. UQpy is a general Python toolbox for uncertainty quantification in engineering and scientific contexts. It implements sampling, propagation, surrogate models, and sensitivity analysis.
Source: github.com/SURGroup/UQpy