Life Cycle Optimization (LCO) with PULPO

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

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

✨ Capabilities

Applying optimization is recommended when the system of study has (1) many degrees of freedom which would prompt the manual assessment of a manifold of scenarios, although only the “optimal” one is of interest and/or (2) any of the following capabilities makes sense within the goal and scope of the study:

• Specify technology and regional choices throughout the entire supply chain (i.e. fore- and background), such as choices for the production technology of electricity or origin of metal resources. Consistently accounting for changes in the background in “large scale” decisions can lead to significantly different insights.

• Specify constraints on any activity in the life cycle inventories, which can be interpreted as tangible limitations such as raw material availability, production capacity, or environmental regulations.

• Optimize and/or constrain any impact category for which the characterization factors are available.

• Specify supply values instead of final demands, which can become relevant if only production values are available (e.g. here).

• Weigh several impact categories against each other, trace Pareto fronts, or set goals for them (goal programming).

• Solve large problems in reduced space (solve(formulation='reduced')): an exact reformulation over the choice alternatives that is much smaller and faster on large databases.

• Optimize over time with PulpoOptimizerTime: demands and limits per time step, impact budgets over the whole horizon, and storage carried from one step to the next.

• Optimize under uncertainty with pulpo.utils.uncertainty: chance-constrained fronts from the declared distributions of the biosphere flows and characterization factors, an exact variance decomposition, and out-of-sample validation.

The examples run on bundled databases with open-source solvers.

💬 Support

If you have any questions or need help, do not hesitate to contact us:

• Fabian Lechtenberg (fabian.lechtenberg@upc.edu)