Abstract: To prevent differential evolution algorithm from being trapped into premature phenomenon， a random mutation differential evolution (RMDE) algorithm was presented. Random choice method was adopted to implement mutation and disturbance operation， which was aimed at increasing the diversity of population and balancing the local and global search. Several standard test functions were tested， and the results verified the RMDE is superior to the other five algorithms. The proposed RMDE algorithm was then applied to solving the economic dispatch problems of power system with 13 units. Experimental results showed that the solution obtained by the RMDE algorithm is better than that reported in recent literatures.

Key words: differential evolution algorithm, random choice, mutation, disturbance, diversity