Corrugated Board Production Planning via Multi-Objective Evolutionary Optimization With Variable Dependency-Based Variation

This study addresses the production planning problem in corrugated board factories and proposes a method to optimize paper selection using multi-objective evolutionary optimization. Each order is represented by a variable that indicates the paper type, and two crossover operators are introduced: flip crossover for paper material and simulated binary crossover for paper width. In addition, orders are clustered based on their dependencies, and the crossover operations are applied at the cluster level, which we call clustered variation. Experiments were conducted using simulated order lists generated based on real factory data. Among several representative multi-objective evolutionary algorithms, NSGA-II showed the best performance. Furthermore, by combining NSGA-II with the proposed clustered variation, better performance and stability in optimization were achieved compared with non-clustered variation and random-clustered variation. In all cases, planning based on evolutionary optimization produced better schedules than those created by human experts. These results demonstrate that the proposed method reduces reliance on expert planners and contributes to automating and improving the long-term efficiency of daily corrugated board production planning.