The primary purpose of this thesis is to propose an algorithm for short-term power scheduling with mixed and limited fuel units.
The underlying model explicitly considers the following factors: setup cost, production cost, ramping rate, operational status, and mixed-limited fuel units. This formulation takes form of mixed integer nonlinear programming problem. Lagrangian Relaxation is used to disaggregate the model by generator type which are then solved by dynamic programming approach.
The strength of the methodology partially lies in its ability to construct good feasible solutions by "Two feasibility steps". Thus, the need for the branch and bound is eliminated. In addition, the model is applied to the Korean electricity sector. Computational experiences with the proposed algorithm indicate that problems of 24 time periods and 23 units with 8 mixed units (approximately comparable to Korean power system configuration) are readily solved within reasonable times.
