Skip to article frontmatterSkip to article content
Site not loading correctly?

This may be due to an incorrect BASE_URL configuration. See the MyST Documentation for reference.

pylcm

pylcm specifies, solves, and simulates finite-horizon discrete-continuous dynamic choice models. You write states, actions, preferences, constraints, and transition laws as ordinary Python functions; pylcm turns that economic description into a backward-induction and simulation pipeline built on JAX.

Use the route that matches what you are trying to do:

The Concepts & Methods chapter explains why the algorithms and representations work. The Development chapter is for people changing pylcm itself.

Scope

pylcm is for finite-horizon models with discrete and continuous states and actions, deterministic or stochastic transitions, and potentially multiple regimes. Brute-force grid search covers the broadest class. Endogenous-grid solvers cover narrower consumption-saving structures and require the model to declare those structures from the outset. GPU acceleration can make large models feasible, but the best solver still depends on candidate growth, memory, compilation, and hardware.

If finite-horizon dynamic programming is new to you, begin with the QuantEcon Dynamic Programming book before using the methods chapter.