An Evolutionary Algorithm for Multi-objective Optimization Problem Based on Projection Plane: MOEA/P
Xiaomeng Lü, Shuang Yang, Funan Peng, Weiru Chen · 2021
Researches on multi-objective optimization pay more attention to the convergence, diversity and robustness of solutions which is oriented to the whole object solution set, large number of scattered solutions affect the efficiency and quality of the algorithms, and too many solutions let user make decision hardly, it is more prominent in many-objective optimization problems. An algorithm of solving the multi-objective optimization problem based on projection plane is proposed in this paper, which divides the objective space into projection plane and free dimension(s) according to the decision demand, carves the projection plane into several projection grids, and solves the optimal value of the free dimension(s) on each projection grid, so as to obtain the optimal solutions of the multi-objective optimization problem. The division of projection plane reduces the problem of high-dimensional multi-objective optimization to that of low-dimensional multi-objective optimization, which only solves the optimization of free dimension objects on the projection plane; The carving of projection grids determines the solutions in the specific range of the specified object values, which improves the accuracy and efficiency of the solutions. The experimental results of multi and many objective optimization show that this algorithm can effectively solve the multi-objective optimization problems, and provide effective support for the decision-making in the specific objects of many-objective optimization problems.