This paper studies the scheduling of power generation in a Microgrid (MG) that has a group of dispatchable and non-dispatchable generators. In order to maximize the benefits of the resources available in a MG, an optimal scheduling of the power generation is required. Renewable resources have an intermittent nature that causes uncertainties in the system. These added uncertainties must be taken into consideration when solving the generation scheduling problem in order to obtain reliable solutions. The operation of a MG in grid-connected mode and isolated mode is analyzed in this paper for different demand profiles. Two mixed integer linear programming (MILP) models for the day-ahead unit commitment problem in a MG are proposed. Each model corresponds to one mode of operation. Uncertainty handling techniques are integrated in both models. The models are solved using the General Algebraic Modeling System (GAMS). Two study cases are examined to study the operation of a MG, and to evaluate the effects of uncertainties on the day-ahead unit commitment problem.