There is not much to do if you mean the R package RobustRankAggreg. You need ranked lists of genes, and then simply feed it into the function. From ?RobustRankAggreg:

glist <- list(sample(letters, 10), sample(letters, 20), sample(letters, 12))
aggregateRanks(glist = glist, N = length(letters))

The lower the score the more often the gene was ranked highly in the individual datasets.