Skip to content

akshayshanker/dynx

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

68 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DynX: Modular Dynamic Programming

⚠️ Experimental Development Build – APIs and documentation are incomplete.⚠️

DynX is a computatioanl framework for representing conjugate pairs of functional operations, a push-forward and pull-back, on Banch space as a graph. The focus of is on a practical rather than theoretical representation of Bellman functional recursions and push-forward measures.

The use of graphs and networks to represent dynamics is ubiquitous (see Auclert et al (2021) or Stachurski and Sargent (2025), for instance). However, the critical innovation in DynX is to represent relations between functions and measures, rather than relations between variables. Moreover, Unlike general graph libraries (e.g. TensorFlow), DynX represents both computational operators and the functional objects that define a recursive problem.


Installation

DynX is not yet published on PyPI. You can install the latest development build directly from GitHub:

Main branch (bleeding-edge)

pip install "git+https://github.com/akshayshanker/dynx.git#egg=dynx"

Specific dev release (v0.18.dev4)

pip install "git+https://github.com/akshayshanker/dynx.git@v0.18.dev4#egg=dynx"

To upgrade an existing installation:

pip install --upgrade --force-reinstall \
    "git+https://github.com/akshayshanker/dynx.git#egg=dynx"

Documentation

Comprehensive documentation is in progress.

About

Modular Dynamic Programming

Resources

License

Stars

Watchers

Forks

Packages

 
 
 

Contributors

Languages