An Indicator and Decomposition Based Steady-State Evolutionary Algorithm for Many-Objective Optimization

Figure 4

The lowest contribution contains more than one individual. Two situations appear. (a) will be eliminated as it belongs to the last contribution level. Meanwhile, there is another better solution associated with the subspace. It is proper for the selection situation. (b) Although belong to the lowest contribution level, it is associated with the subspaces . Simply adopting the indicator is improper. Thus, we eliminate the most crowded subspace solution in through adopting the PBI decomposition method. Achieving balance between convergence and diversity is mainly concerned through the indicator and decomposition method.

(a) Proper situation

(b) Improper situation