Generating Random Permutations in the Framework of Parallel Coarse Grained Models

We present three algorithms for generating random permutations in the coarse grained model CGM. For each of the proposed algorithms, we study the number of supersteps, the size of the local memory, the overall communicat- ion cost and we check if it gives a permutation with the uniform distribution or not. The proposed algorithms are intended to be simple and of practical relevance. The difficulty, in this paper, lies in proving that they are the desired properties.