Considering multiple factors such as production efficiency, cost-effectiveness, and environmental sustainability, this paper comprehensively investigates the correlations among eight common optimization objectives in the Integrated Process Planning and Scheduling (IPPS) problem. An evolutionary algorithm based on objective dimensionality reduction is proposed to solve this Many-objective IPPS (MaO-IPPS) problem. First, a hybrid initialization method is designed by combining multiple scheduling strategies, which is conducive to generating a high-quality and diverse initial population. Next, the objective dimensions of the problem are reduced by eliminating redundant objectives and aggregating objectives with high harmony, thereby reducing problem complexity. Then, Ensemble Fitness Ranking (EFR) is introduced to construct a novel dominance relation for evaluating and selecting individuals, and the idea of hierarchical iteration is applied to guide the algorithm to converge quickly while maintaining solution diversity. Finally, the proposed algorithm is experimentally compared with other algorithms such as MOEA/D and NSGA-III on 24 instances of different sizes, and its performance is further evaluated by using metrics such as Hyper Volume (HV) and Inverted Generational Distance (IGD). The experimental results show that the proposed algorithm possesses better comprehensive performance and can obtain better Pareto solutions than the comparison algorithms, especially as the problem scale and the number of objectives increase.
Yang et al. (Thu,) studied this question.