Cataclysmic Genetic Algorithm for Reactive Power Dispatch Comparing with an Interior-Point Algorithm

Y.-J. Zhang, M.-Y. Liao, and Z. Ren (PRC)


Genetic algorithms, reactive power dispatch, and optimization


The optimal reactive power dispatch (ORPD) problem of power systems is one of the mixed-integer nonlinear optimization problems. Genetic algorithms (GAs) are used to solve this kittle problem since it can search for a global optimum using multiple path and treat discrete problems naturally. However, premature convergence limits its application to real-time reactive power control. Mimicking the cataclysm of the evolution course of eco-systems, in which most species are extinct but very few survive, the cataclysm operator updates all individuals randomly except for the current optimum after tens of generations. With some other improving measures, cataclysmic genetic algorithm (CGA) can therefore enhance population diversity and overcome the insufficiency of GAs. The numerical test based on IEEE 118-bus system comparing with an interior-point algorithm is presented and demonstrate the proposed CGA can obtain better results.

Important Links:

Go Back