Nhniina bound cyclic

Constraint q_boundCyclic wasn't working correctly in the case of several samples that were meant to be in series. There already existed a parameter p_msWeight which apparently wasn't used anywhere in the code. It is now used to give different temporal weights for different samples. Should most likely be one unless using samples as representative periods.

Please check that the updates don't break things in long-term scenario simulations or in building/schedule runs with multiple forecasts.