A HEURISTIC SHORT-TERM UNIT COMMITMENT

This paper considers a heuristic approach to the unit commitment problem. The proposed algorithm uses priority list based heuristics in the form of inference rules to find a suboptimal schedule for a given load pattern. The new schedule is a modified and locally optimized version of an existing schedule that is retrieved from the system database. The global optimization can be performed as an inference process once the suboptimal solution is obtained. A prototype system is developed and tested on an IBM-PC under EASE+NEXPERT shell and the C language. The performance of the system demonstrates the real-time processing capability and the effectiveness of the proposed system, particularly for the short-term unit commitment.