A combined GA and IPM approach for unit commitment problem

Unit commitment involves determination of a large number of variables (both discrete and continuous). The well-known curse of dimensionality is very pronounced when dealing with large-scale power systems. This paper presents an approach combined between genetic algorithm (GA) and interior point method (IPM). GA is used to determine the on/off status of generators (discrete variables) whereby IPM finds the optimal power dispatch of the committed generators. Simulations were carried out for the case with 10, 30, 60 and 100 units. The results show that the proposed method helps improve quality of the final result and also reduce computing time.