Input: |
NP: the population size. |
Max_FES: maximum number of function evaluation. |
Ω: the oracle pamameter. |
the generation strategies pool: “rand/1/bin”, “rand/2/bin”, and “current-to-rand/1”. |
the control parameters pool: [, = 0.1], [], and []. |
(1) ; |
(2) if there are discrete variables, then the generalized discrete variables handling method is used, |
(3) generate an initial population by uniformly and randomly sampling from the search landspace, |
(4) measure the constraint violations with the residual function for each individual in current population, |
(5) evaluate the modified oracle penatly function values for each individual in current population, |
(6) FES = NP; |
(7) while FES < Max_FES do |
(8) ; |
(9) for : NP do |
(10) use the three strategies, with a control parameter setting randomly selected from the control parameter pool, to generate three |
trial vectors and for the target vector ; |
(11) measure the constraint violations with the residual function values of the three trial vectors and ; |
(12) evaluate the modified oracle penatly function values of the three trial vectors and ; |
(13) choose the best trial vector and add it into ; |
(14) FES = FES + 3; |
(15) end for |
(16) ; |
(17) end while |
(18) select the individual with the smallest modified oracle penatly function value in the population |
Output: the objective function value |