An Evolutionary Many-Objective Optimization Algorithm Based on IGD Indicator and Region Decomposition
Shuifeng Feng, Jiechang Wen · 2019
We propose a many-objective optimization algorithm based on IGD indicator and region decomposition in this paper. We divide the objective space into multiple sub-regions by a set of direction vectors. Individuals in each sub-region are compared independently. There are some individuals in each sub-region, which can improve the diversity of algorithm. In the selection of operator, we select individuals according to rank first, and then we design different selection methods in different ranks to select individuals with good performance. In the experiment, we compare the proposed algorithm with MaOEA/IGD in 12 test functions with 10 to 20 objectives, and the results show that the proposed algorithm is effective.