We introduce a stochastic model for the problem of scheduling the generating units of electric utilities. The uncertainty is modeled by choosing a set of possible scenarios. Each scenario is assigned a probability that reflects its likelihood of occurrence. Numerical results obtained by applying this model to the power generation problems of MEPCC indicate that we can achieve substantial savings in the operating cost when the stochastic model is used. Due to the separable nature of the model, the problem can be solved using parallel machines.