R version of "Two efficient algorithms to find Pareto Front" from Matlab Central http://www.mathworks.com/matlabcentral/fileexchange/17251-pareto-front
install.packages("Rcpp")
source("paretoFront.R")
## Identify pareto front of random points
## Returns logical vector of same number of rows as input dataset
M=matrix(runif(1e5),ncol=2)
pp <- paretofront(M)
plot(M[pp,])
## Same, but splits the given objective set into several smaller groups
## to be examined by the first algorithm, for potential further efficiency gains
pp<- paretoGroup(M)