A hybrid Univariate Marginal Distribution Algorithm for dynamic economic dispatch of units considering valve-point effects and ramp rates

This paper presents a new approach for dynamic economic dispatch (DED) problem in power system by using a hybrid Univariate Marginal Distribution Algorithm (HUMDA). The DED problem with valve-point effects and ramp rate limits is a nonliner constrained optimization problem with non-convex and non-smooth characteristics. In the proposed method, a two-stage adaptive mechanism is devised to control parameters of the Univariate Marginal Distribution Algorithm in continuous domains (UMDAc) dynamically and lead the algorithm with better search efficiency; a chaotic local search operator is integrated with UMDAc to effectively avoid premature convergence. Moreover, a constraint handle according to the two-stage adaptive mechanism is proposed, and the results show that the strategy can handle constraints effectively. Finally, the efficiency of the proposed method is validated on two test systems consisting of 5, 10 and 30 thermal units. The results show the superiority of the proposed method while it is compared with other works in the area. Copyright (c) 2013 John Wiley & Sons, Ltd.