Heterogeneous Differential Evolution for Numerical Optimization

Algorithm 1

The dynamic heterogeneous DE (dHDE).

(1) Randomly initialize the population 𝑃 including the variables and ID;

(2) Evaluate the objective fitness value of each individuals in 𝑃;

(3) FEs =  𝑁_𝑝;

(4) while FEs ≤ MAX_FEs do

(5)   for  𝑖=1 to 𝑁_𝑝  do

(6)    if  ID_𝑖==1  then

(7)       Use DE/rand/1/bin to generate a trail vector 𝑈_𝑖;

(8)    end

(9)    if  ID_𝑖==2  then

(10)       Use DE/best/1/bin to generate a trail vector 𝑈_𝑖;

(11)    end

(12)    if  ID_𝑖==3  then

(13)       Use DE/BoR/1/bin to generate a trail vector 𝑈_𝑖;

(14)    end

(15)    Evaluate the objective fitness value of 𝑈_𝑖;

(16)    FEs++;

(17)    if f 𝑈_𝑖 ≤   f 𝑋_𝑖  then

(18)        𝑋_𝑖=𝑈_𝑖

(19)    end

(20)    else

         /* For static HDE (sHDE), please delete lines 20−−22  */

(21)       Randomly assign a new different DE scheme to 𝑋_𝑖, (change the value of ID_𝑖);

(22)    end

(23)    end

(24) end