The parametric study and adaptive handling improvement of differential evolution algorithm
Wang Xin-dan · Journal of Shandong Jianzhu University · 2006
In recent years,a new method called differential evolution(DE) algorithm has been developed.This method shows many advantages: faster convergence,less control variables,and easier being put into practice,etc.The detailed information about the basic theory of DE is given,and for fastening evolutional speed,the choice of the DE key parameters,such as mutation factor F and crossover factor CR,is presented through four classical De Jong functions.Then an adaptive mutation differential evolution(AMDE) algorithm is proposed.Some tests demonstrate that AMDE is faster than DE.And it needs only two key parameters,greatly reducing sensitivity to parameters,thus simplifying its operations.