dcegm

<!-- Python implementation of the Discrete-Continuous Endogenous Grid Method (DC-EGM) for solving dynamic stochastic lifecycle models of continuous (e.g. consumption-savings) and additional discrete choices. -->

Note: This is a pre-release version of the package. While the core features are in place, the interface and functionality may still evolve. Feedback and contributions are welcome.

dcegm is a Python package for solving and simulating finite-horizon stochastic discrete-continuous dynamic choice models using the DC-EGM algorithm from Iskhakov, Jørgensen, Rust, and Schjerning (QE, 2017).

The solution algorithm employs an extension of the Fast Upper-Envelope Scan (FUES) from Dobrescu & Shanker (2022).

Installation

You can install dcegm via PyPI or directly from GitHub. In your terminal, run:

$ pip install dcegm

To install the latest development version directly from the GitHub repository, run:

$ pip install git+https://github.com/OpenSourceEconomics/dcegm.git

Documentation

The documentation is hosted at https://dcegm.readthedocs.io

References

1. Christopher D. Carroll (2006). The method of endogenous gridpoints for solving dynamic stochastic optimization problems. Economics Letters
2. Iskhakov, Jorgensen, Rust, & Schjerning (2017). The Endogenous Grid Method for Discrete-Continuous Dynamic Choice Models with (or without) Taste Shocks. Quantitative Economics
3. Loretti I. Dobrescu & Akshay Shanker (2022). Fast Upper-Envelope Scan for Discrete-Continuous Dynamic Programming.

<!-- ## Citation If you use dcegm for your research, please do not forget to cite it. ```bibtex @Unpublished{BleschGsell2025, Title = {dcegm: A GPU-accelerated python implementation of the Discrete-Contiunous Endogenous Grid Method}, Author = {Maximilian Blesch & Sebastian Gsell}, Year = {2025}, Url = {https://github.com/OpenSourceEconomics/dcegm} } ``` -->