README.md

April 24, 2026 · View on GitHub

Python-based User-defined Lifecycle Production Optimization

Jupyter Python Markdown

License Last Commit Commit Activity

PyPI - Version GitHub Stars launch - renku

📍 Overview

PULPO is a Python package for Life Cycle Optimization (LCO) based on life cycle inventories. It is designed to serve as a platform for optimization tasks of varying complexity.

The package builds on top of the Brightway LCA framework and the Pyomo optimization modeling framework.

✨ Capabilities

Applying optimization is recommended when the system of study has (1) many degrees of freedom that would otherwise prompt the manual assessment of a large number of scenarios, or (2) any of the following capabilities is relevant to the goal and scope of the study:

  • Specify technology and regional choices throughout the entire supply chain (fore- and background), such as the production technology of electricity or the origin of metal resources. Consistently accounting for background changes in large-scale decisions can be significant.
  • Specify constraints on any activity in the life cycle inventories, interpreted as tangible limitations such as raw material availability, production capacity, or environmental regulations.
  • Optimize for or constrain any impact category for which characterization factors are available.
  • Specify supply values instead of final demands, which is relevant when only production volumes are known (e.g. here).
  • Optimize under uncertainty via a dedicated pipeline: import and filter uncertain LCI parameters, apply uncertainty strategies, run Global Sensitivity Analysis (Sobol), perform Monte Carlo sampling, or solve Chance-Constrained programs to obtain Pareto-optimal solutions at user-defined probability levels.

Features recently completed:

  • ℹ️ Optimization under uncertainty [chance-constraints, Monte Carlo, global sensitivity analysis]
  • ℹ️ Development of a GUI for simple optimization tasks Link
  • ℹ️ Enable PULPO to work on both bw2 and bw25 projects
  • ℹ️ Thorough documentation hosted on flechtenberg.github.io/pulpo/

Features currently under development:

  • ℹ️ Multi-objective optimization [bi-objective epsilon constrained, goal programming ...]
  • ℹ️ Integration of economic and social indicators in the optimization problem formulation

Feature requests are more than welcome!

🔧 Installation

PULPO is available on PyPI. Depending on the version of Brightway you want to work with, install either the bw2 or bw25 variant:

pip install "pulpo-dev[bw2]"

or

pip install "pulpo-dev[bw25]"

🤖 Running PULPO

Use this link to start a cloud session and test PULPO right away:

launch - renku

The main reference is the PULPO showcase notebook, which revolves around methanol production and covers both the core optimization features (Sections 1–9) and the full workflow of the pulpo_unc module (Section 10): uncertainty data import and filtering, gap-filling strategies, Monte Carlo from prepared distributions, Chance-Constrained optimization, and Global Sensitivity Analysis.

Additional example notebooks are available for a hydrogen case, an electricity case, and a plastic case.

There is also a workshop repository (here) created for the Brightcon 2024 conference, with guided notebooks and exercises.

🧪 Tests

Run from the package folder:

python -m unittest discover -s tests

What's new in 1.5.1?

  • Hotfix for chance-constrained uncertainty workflow: corrected environmental-cost mean computation in pulpo.utils.uncertainty.cc by replacing fragile pandas sparse updates with direct sparse-matrix updates and multiplication.
  • Updated Section 10 of the main showcase to include the deterministic reference result used by run_gsa, and re-ran the workflow end-to-end.
  • Moved the sample and rice-husk database helpers into a new pulpo.datasets subpackage so that pulpo.install_sample_db() and pulpo.install_rice_husk_db() work from an installed wheel.
  • Packaging fixes: tests/ is no longer included in the wheel, and the bw2 / bw25 extras are now declared as mutually exclusive in [tool.uv] conflicts so that uv can resolve the lockfile cleanly.

What's new in 1.5.0?

This release integrates the full uncertainty analysis pipeline into PULPO via the new pulpo.pulpo_unc module, turning the long-running development effort into a first-class feature:

  • PulpoOptimizerUnc — A subclass of PulpoOptimizer that exposes the entire uncertainty workflow through a single worker object.
  • Uncertainty data import and filtering — Import uncertain parameters directly from Brightway databases via import_and_filter_uncertainty_data(), with configurable cutoff-based filtering.
  • Uncertainty strategies — Fill missing or incomplete uncertainty specifications using apply_uncertainty_strategies(), with built-in strategies (e.g. triangular bound interpolation) and support for custom expert-knowledge distributions.
  • Monte Carlo from prepared distributions — Run MC on the curated uncertainty data (without re-sampling the full Brightway matrices) via run_mc_from_uncertainty().
  • Chance-Constrained (CC) optimization — Formulate and solve chance-constrained programs with create_CC_formulation() and solve_CC_problem(), yielding a Pareto front of optimal solutions at varying probability (risk) levels.
  • Global Sensitivity Analysis (GSA) — Identify the uncertain parameters that drive optimization outcomes using Sobol sensitivity indices via run_gsa().
  • End-to-end showcase — Section 10 of the PULPO showcase notebook walks through the complete pulpo_unc workflow on the methanol system.
  • Note: The uncertainty features in this release have been implemented and tested with Brightway bw2 only; users working with bw25 should validate workflows and may need to adapt configuration.
  • Minor bugfixes and code cleanup.

What's new in 1.4.3?

  • Allow users to pass lower inventory flow and lower impact limits via lower_inv_limit and lower_imp_limit dicts.
  • Provide new showcase notebook.
  • Enable users to pass custom default upper limits on elements, given that Gurobi identified 1e20 (and 1e24) as infinite in some cases. See section 8 of the showcase for usage. Setting them lower may also improve convergence speed.
  • Enable dependent constraint definition. See section 9 of the showcase for usage.

What's new in 1.4.2?

  • Enable the use of Gurobi solver.

What's new in 1.4.0?

  • Enable the use of NEOS solver (commercial solvers without a local license).
  • Enable Monte Carlo sampling feature.
  • Retrieve uncertainty information to lci_data for future use.

What's new in 1.3.0?

  • Switch packaging logic from setup.py to pyproject.toml and align PyPI with GitHub versioning.

🤝 Contributing

Contributions are very welcome. To request a feature or report a bug, please open an Issue. If you are confident in your coding skills, feel free to implement your suggestions and send a Pull Request.

📄 License

This project is licensed under the ℹ️ BSD 3-Clause License. See the LICENSE file for additional info.
Copyright (c) 2026, Fabian Lechtenberg. All rights reserved.

👏 Acknowledgments

We would like to express our gratitude to the authors and contributors of the following packages that PULPO builds upon:

We also acknowledge the pioneering ideas and contributions from the following works:

The development of PULPO culminated in the following publication, which details the approach and outlines its implementation:

Fabian Lechtenberg, Robert Istrate, Victor Tulus, Antonio Espuña, Moisès Graells, and Gonzalo Guillén‐Gosálbez.
“PULPO: A Framework for Efficient Integration of Life Cycle Inventory Models into Life Cycle Product Optimization.”
Journal of Industrial Ecology, October 10, 2024.
https://doi.org/10.1111/jiec.13561

Please cite this article if PULPO is used to produce results for a publication or project.

Authors

↑ Return