policyengine-simulation-mechanics

Use Simulation.ensure() to enable smart caching that reuses baseline results across reforms. The Skill implements in-memory LRU caching for up to 100 simulations and disk persistence via save/load, eliminating redundant computation in iterative policy analysis workflows.

Access results via output_dataset.data to retrieve simulation outputs, then use map_to_entity() to convert data between person, household, and other entity levels. This lets you analyze outcomes at the appropriate granularity for your policy question.

Caching is essential for batch scenario simulation and comparative policy analysis where you reuse the same baseline across multiple reforms. It eliminates inefficient recomputation and dramatically speeds up iterative development and what-if analysis.

Yes. Beyond in-memory LRU caching, the Skill provides disk persistence via save/load methods, letting you preserve simulation state across sessions and share cached results for reproducible policy analysis.

The map_to_entity() function converts data between person, household, and benunit/entity levels, enabling you to aggregate or disaggregate simulation outputs to match your analytical or reporting requirements.

Create a baseline Simulation, call ensure() to cache it, then run reforms against that cached baseline. This workflow maximizes reuse of computation across multiple policy scenarios and minimizes redundant processing.