Skip to content

CVXR v1.8.1: A completely new rewrite for Convex Programming in R

1 message · Marc Schwartz via R-packages

#
Dear useRs,

We are happy to announce the release of CVXR 1.8.1 on CRAN, unleashing 
the full power of convex optimization for R users. Version 1.8.1 is a 
ground-up rewrite using S7 classes and establishes parity with CVXPY 
1.8.1, except for some esoteric atoms that are rarely needed. We support 
four open source solvers (osqp, clarabel, scs and highs) out of the box; 
additional solvers, open source and commercial, are also supported. CVXR 
also supports Disciplined Parametric Programming (DPP), Disciplined 
Geometric Programming (DGP) and DQCP (Disciplined Quasiconvex Programming).

We also note that people are often surprised to hear that one can 
formulate Integer Programming problems with CVXR.

The full list of features is large, so please consult the website

https://cvxr.rbind.io

for what's new, migration guide, and a number of examples. A totally new 
feature in CVXR is visualization: see `?visualize` and examples on the 
website.

A related talk regarding the development of this new version may be of 
interest. Slides and video at

https://naras.su.domains/blog/ (Slide 25 for TL;DR.)

We welcome your feedback and issues on Github: 
https://github.com/cvxgrp/CVXR

Thank you.

Anqi Fu and Balasubramanian Narasimhan

_______________________________________________
R-packages mailing list
R-packages at r-project.org
https://stat.ethz.ch/mailman/listinfo/r-packages