To set the record, it is universally understood that "an" optimum moisture content is energy driven. In other words, OMCvp is higher (likely) than OMCstd which, in turn, is higher than OMCmod which is greater than OMC15tvib. It is obvious that 95% compaction relates to either Standard Proctor (D698) or Modified Proctor (D1557) (or, for the Brits, light tamping and heavy tamping respectively). 95% compaction, then can be achieved over a range of moisture contents (compared to the standard or modified proctors). Obviously, for your plate compactor you should be attempting to achieve "optimum" plate which isn't the reference standard optimum - as this will give you the maximum density that can be achieved. You will need to confirm that the OMCvp is within the range of the moisture contents that, when achieved, can produce something which gives a dry density above your limits. This likely is achievable. The need to minimize the loose lift thicknesses is to be able to impart as much energy as possible into the material - the thicker the lift the less energy throughout the whole lift. Too, there is also the fact that overlying lifts will cause additional compaction in lower lifts - although with vibrating plate, not much. As fattdad says, a trial pad is the only way to be sure - but not, in my view, to determine necessarily the optimum moisture content but that the method will provide 95% of the referenced compaction at the moisture content you have - if not, then you would have to dry or wet the soil to see if you can achieve better compaction - or adjust the lift thickness.