I am trying use the rankprod package but i am unable to do so. Its like initially I had .txt files and I think RankProd does not takes .txt format. Please tell me if it does and how shall I proceed? Also if I take data in .cel format then I do not have the sample-description.csv file. So what shall I do?

If you have celfiles, then you should use either the affy or oligo or xps package to normalize and summarize your data. You then need a 'classlabel' vector of 0s and 1s, indicating which group a given sample comes from. RankProd won't accept an ExpressionSet, so you have to extract it. As an example, say you have six samples, three of one type and three of another. Also assume these are Affy Gene 2.0 ST arrays. This is how I would analyze the data.

library(oligo)
library(RankProd)
dat <- read.celfiles(filenames = list.celfiles())
eset <- rma(dat)
cl <- rep(0:1, each  = 3) ## assuming that the first three and last three are different
results <- RP(exprs(eset), cl)