Research Article

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

Figure 6

The illustration of DBEA and MOEA/DD. (a) Selection strategy based on DBEA. are the combined solution. are a set of reference vectors. If a new individual is generated, it will be deleted from the combined population. The subspace will be discarded. The convergence is well while the diversity will be damaged. (b) Selection strategy based on MOEA/DD. If a new individual is generated, a sophisticated steady-state Pareto ranking strategy ENLU is used to update the rank. The PBI decomposition method is used to select the proper individuals. The star points will be survived. Combining Pareto dominance and decomposition method can well balance convergence and diversity.
(a) Selection based on DBEA
(b) Selection based on MOEA/DD