Optimizing Assembly Sequence Planning Using Precedence Graph-Based Assembly Subsets Prediction Method

Nan Zhang,Zhenyu Liu,Chan Qiu,Weifei Hu,Jianrong Tan

DOI: https://doi.org/10.1108/aa-02-2019-0031

IF: 2.373

2020-01-01

Assembly Automation

Abstract:Purpose Assembly sequence planning (ASP) plays a vital role in assembly process because it directly influences the feasibility, cost and time of the assembly process. The purpose of this study is to solve ASP problem more efficiently than current algorithms. Design/methodology/approach A novel assembly subsets prediction method based on precedence graph is proposed to solve the ASP problem. The proposed method adopts the idea of local to whole and integrates a simplified firework algorithm. First, assembly subsets are generated as initial fireworks. Then, each firework explodes to several sparks with higher-level assembly subsets and new fireworks are selected for next generation according to selection strategy. Finally, iterating the algorithm until complete and feasible solutions are generated. Findings The proposed method performs better in comparison with state-of-the-art algorithms because of the balance of exploration (fireworks) and exploitation (sparks). The size of initial fireworks population determines the diversity of the solution, so assembly subsets prediction method based on precedence graph (ASPM-PG) can explore the solution space. The size of sparks controls the exploitation ability of ASPM-PG; with more sparks, the direction of a specific firework can be adequately exploited. Practical implications The proposed method is with simple structure and high efficiency. It is anticipated that using the proposed method can effectively improve the efficiency of ASP and reduce computing cost for industrial applications. Originality/value The proposed method finds the optimal sequence in the construction process of assembly sequence rather than adjusting order of a complete assembly sequence in traditional methods. Moreover, a simplified firework algorithm with new operators is introduced. Two basic size parameters are also analyzed to explain the proposed method.