There are currently a number of R packages available for solving optimization problems, either using proprietary solvers or with built-in solvers. In addition, many of those packages provide tools for reading a pre-specified problem from a file, but they are limited to linear and mixed-integer problems.
ReadQP aims to build on top of that functionality by providing the ability to also handle quadratic programming problems by wrapping the Rglpk_read_file function from Rglpk and adding additional steps to parse a quadratic objective into an in-memory representation that is compatible with ROI.
Handle all variations of the model goal (e.g. max, min, maximize, maximum, upper and lower case etc.)Handle all variations of the constraint statement (Subject To, s.t., such that, ST., etc.)- Nested quadratic components in the objective
To install readQP from GitHub, you will need the devtools package installed. Instructions for installing devtools can be found here.
Once you have devtools installed, simply run the following command from your R console to install readQP:
library(devtools)
devtools::install_github('cafreeman/readQP')To verify the installation, you can run the test suite with the following command:
library(testthat)
library(readQP)
testthat::test_package('readQP')readQP currently exposes one main function, readModelFile, which takes a file path and a file type as arguments and returns an in-memory representation of the model.
To see a basic example, execute the following commands:
sampleFile <- readQP:::getSampleData("simple_qp")
qpModel <- readModelFile(sampleFile)
str(qpModel)The resulting qpModel object will contain all the various components of the model, including the quadratic component which can be found at qpModel$objective$Q.