Application of stochastic programming and probabilistic analyses as key parameters for real decision making regarding implementing or not energy rationing - a case study for the Brazilian hydrothermal interconnected system.