Solves weighted bipartite matching problems (e.g., optimal matching of people to cars
or optimal matching of students to colleges, etc...)
Usage
HungarianSolver(costMatrix)
Value
List with cost and parings, pairings are given as an Nx2 matrix
giving edges that are matched (1-indexed rather than 0-indexed as it will be returned to R)
Arguments
costMatrix
matrix giving cost of each possible pairing - can be rectangular
Details
this is a copy/wrapper for the code developed by Cong Ma and made available
as a github repository (mcximing/hungarian-algorithm-cpp). Code was
changed to a header only file for use in other Rcpp packages.