An Application of Mixed-Integer Programming Duality to Scheduling Thermal Generating Systems

A decomposable mixed-integer programming model for simultaneous economic consideration of unit commitment and short-term dispatch of thermal power generating equipment is presented. An optimum solution procedure is described which takes advantage of the structure of the model and promises computational efficiency in practical applications. Characterizing the demand forecast as a discretized function permits a probabilistic forecast to be incorporated in the scheduling model. The model minimizes the total expected costs of start-up, shutdown, and unit operating costs subject to the demand, reserve, and unit commitment constraints. Extensions to multiple periods and to include additional constraints such as time dependent start-up costs and geographic constraints are given.