Disciplined Convex Programming in R using 'Convex.jl'

Provides a simple high-level wrapper for 'Julia' package 'Convex.jl' (see < https://github.com/JuliaOpt/Convex.jl> for more information), which makes it easy to describe and solve convex optimization problems in R. The problems can be dealt with include: linear programs, second-order cone programs, semidefinite programs, exponential cone programs.

Travis-CI BuildStatus AppVeyor BuildStatus CRAN_Status_Badge

convexjlr is an R package for Disciplined Convex Programming (DCP) by providing a high level wrapper for Julia package Convex.jl. The aim is to provide optimization results rapidly and reliably in R once you formulate your problem as a convex problem. convexjlr can solve linear programs, second order cone programs, semidefinite programs, exponential cone programs, mixed-integer linear programs, and some other DCP-compliant convex programs through Convex.jl.


convexjlr is on CRAN now! To use package convexjlr, you first have to install Julia https://julialang.org/ on your computer, and then you can install convexjlr just like any other R packages.

Note: convexjlr used to support multiple ways to connect to julia, one way was through package XRJulia and the other way was to use package JuliaCall. The latter approach was more performant and thus the default approach. But due to the fact that XRJulia doesn’t support julia v0.7 and v1.0 yet, only JuliaCall backend is supported currently.

We hope you use convexjlr to solve your own problems. If you would like to share your experience on using convexjlr or have any questions about convexjlr, don’t hesitate to contact me: [email protected].

Quick Example

We will show a short example for convexjlr in solving linear regression problem. To use package convexjlr, we first need to attach it and do the initial setup:

#> Attaching package: 'convexjlr'
#> The following object is masked from 'package:base':
#>     norm
## If you wish to use JuliaCall backend for performance
convex_setup(backend = "JuliaCall")
#> Doing initialization. It may take some time. Please wait.
#> Julia version 1.0.2 at location /Applications/Julia-1.0.app/Contents/Resources/julia/bin will be used.
#> Loading setup script for JuliaCall...
#> Finish loading setup script for JuliaCall.
#> [1] TRUE

And this is our linear regression function using convexjlr:

linear_regression <- function(x, y){
    p <- ncol(x)
    ## n is a scalar, you don't have to use J(.) to send it to Julia.
    n <- nrow(x) ## n <- J(nrow(x))
    ## x is a matrix and y is a vector, you have to use J(.) to send them to Julia.
    x <- J(x)
    y <- J(y)
    ## coefficient vector beta and intercept b.
    beta <- Variable(p)
    b <- Variable()
    ## MSE is mean square error.
    MSE <- Expr(sumsquares(y - x %*% beta - b) / n)
    ## In linear regression, we want to minimize MSE.
    p1 <- minimize(MSE)
    list(coef = value(beta), intercept = value(b))

In the function, x is the predictor matrix, y is the response we have. And the linear_regression function will return the coefficient and intercept solved by cvx_optim.

Now we can see a little example using the linear_regression function we have just built.

n <- 1000
p <- 5
## Sigma, the covariance matrix of x, is of AR-1 strcture.
Sigma <- outer(1:p, 1:p, function(i, j) 0.5 ^ abs(i - j))
x <- matrix(rnorm(n * p), n, p) %*% chol(Sigma)
## The real coefficient is all zero except the first, second and fourth elements.
beta0 <- c(5, 1, 0, 2, 0)
y <- x %*% beta0 + 0.2 * rnorm(n)
linear_regression(x, y)$coef
#>              [,1]
#> [1,]  5.003248799
#> [2,]  0.991593361
#> [3,] -0.013118929
#> [4,]  2.008255127
#> [5,]  0.004305963

More Examples

More examples (including using convexjlr for Lasso, logistic regression and Support Vector Machine) can be found in the pakage vignette or on the github page: https://github.com/Non-Contradiction/convexjlr



  • Updates for Julia v0.7 and v1.0.
  • Drop XRJulia support, as it does not work with Julia v0.7 and v1.0.


  • Default SCS solver doesn't have verbose = FALSE default option any more.
  • Users can choose ECOS as the solver for convex problems.
  • Users can set a bunch of options for both SCS and ECOS solvers.


  • The users can set maximal iteration times for the convex problem solver in cvx_optim.
  • Bug correction for handling of diag.

convexjlr 0.7.0

  • Remove deprecated setup function.
  • Use JuliaCall as the default backend.


  • Fix deprecation warnings from JuliaCall backend.
  • Fix some little bugs.
  • Add the option in convex_setup to set the path to julia binary.

convexjlr 0.6.1

  • The second release on CRAN.


  • Supports multiple ways to connect to julia, one way is through package XRJulia, and the other way is to use package JuliaCall. The difference is as follows:
    • XRJulia connects to julia, which is the default for convexjlr, the advantage is the simplicity of the installation process, once you have a working R and working julia, it should be okay to use convexjlr in this way. Note that if you have the latest Julia version (v0.6.0) installed, then you have to use the latest version of XRJulia.
    • JuliaCall embeds julia in R, the advantage is the performance, for example, if your convex problem involves large matrice or long vectors, you may wish to use JuliaCall backend for convexjlr; the disadvantage is the installation process, since embedding julia needs compilations.


  • Added a NEWS.md file to track changes to the package.
  • Re-organize tests.
  • Deprecate setup, should use convex_setup.

convexjlr 0.5.1

  • A patch release on CRAN.

convexjlr 0.5.0

  • The first release on CRAN.

Reference manual

It appears you don't have a PDF plugin for this browser. You can click here to download the reference manual.


0.8.1 by Changcheng Li, 4 months ago


Browse source code at https://github.com/cran/convexjlr

Authors: Changcheng Li [aut, cre]

Documentation:   PDF Manual  

Apache License | file LICENSE license

Imports JuliaCall, magrittr

Suggests testthat, knitr, rmarkdown, plotrix

System requirements: Julia (>= 0.6.0), Convex.jl, SCS.jl, ECOS.jl

See at CRAN