Local optimal search algorithm for unit commitment problem

The unit commitment problem (UCP) is an optimization problem that decides start-up time, downtime and electricity generation amount that meets the given constraints for the multiple generators of a power system. Establishing an efficient unit commitment plan has recently become important for economic operation due to increasing power demand and the variety of generator types. The formulation of unit commitment considers the constraints such as start-up cost, minimum start-up time, and minimum downtime. Due to the structural features of the constraints, an on-off schedule is decided first, followed by the generation for each power unit when solving the UCP problem. In this paper, we discuss an optimal generation distribution after the on-off schedule is fixed. First, we derive an optimal condition for the generation distribution with fixed schedule, and then propose an efficient method to search for the optimal generation if a given generation distribution is not an optimal solution. Finally, we propose a heuristic method of searching for the adjacent schedule for the feasible solution of UCP with fixed generation schedule. Using the proposed methods, we evaluate whether the solutions proposed in previous studies were local optimal solutions, and, if not, search for new local optimal solutions.