A Modified Two_Arch2 Based on Reference Points for Many-Objective Optimization

Many-objective optimization problems (MaOPs) refer to those multi-objective problems (MOPs) having more than three objectives. For MaOPs, the performance of most multi-objective evolutionary algorithms (MOEAs) often deteriorates because it is hardly to achieve a balance between convergence and population diversity. To address this issue, this paper proposes a modified Two_Arch2 based on reference points (called Two_Arch2-RP). Firstly, a simplified reference point based dominance (RP-dominance) and cosine distance are used to update the convergence archive (CA). Then, the niche-preserving operation based reference points is combined with Pareto dominance to maintain the diversity archive (DA). To test the optimization capability of the proposed Two_Arch2-RP, seven DTLZ benchmark problems with 3, 5, 8, and 15 objectives are used. Experimental results show that Two_Arch2-RP obtains superior performance when compared with five other state-of-the-art algorithms.