Add documentation specifying random seeds for parallel problems

[PrasadBabarendaGamage]

For multi-processing examples, we need to run the random seed in the example - that will make sure the partitioning is consistent because Parmetis is heuristic and may change. Best to include this in the example creation documentation.