CVXR provides an object-oriented modeling language for convexoptimization, similar toCVX,CVXPY,YALMIP, andConvex.jl. It allows the user toformulate convex optimization problems in a natural mathematical syntaxrather than the restrictive standard form required by most solvers. Theuser specifies an objective and set of constraints by combiningconstants, variables, and parameters using a library of functions withknown mathematical properties.CVXR then applies signeddisciplinedconvex programming (DCP) to verify the problem’s convexity. Onceverified, the problem is converted into standard conic form using graphimplementations and passed to a cone solver such asECOS orSCS.
CVXR includes several open source solvers in addition to the defaultOSQP, ECOS and SCS. Recent (1.x+) versions also include support forcommercial solvers such asMOSEK,GUROBI andCPLEX.
For details and examples, we refer you toFu, Narasimhan, Boyd(2020). If you use CVXR in your work, please cite this reference. (The Rcommandcitation("CVXR", bibtex = TRUE) will also give youa bibtex-formatted entry.)
This package is now released on CRAN, so you can install the currentreleased version as you would any other package for R, version 3.4 andhigher. (CVXR is known to work with earlier versions of Rtoo, but we don’t check our releases against older versions of R.)
install.packages('CVXR',repos ="https://CRAN.R-project.org")Development versions can be installed from the Github repositoryassuming you have the development tools for R available, including the Ccompilers etc. Execute:
library(devtools)install_github("cvxgrp/CVXR")A number of tutorial examples are available on theCVXR website along with links to ouruseR! 2019 short-course.