Optimizing assembly sequence planning using precedence graph-based assembly subsets prediction method

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.

Digital twin-based multi-level task rescheduling for robotic assembly line

“…To obtain a scheduling plan with high solution quality in the shortest possible computing time can meet the requirements of near real-time scheduling in DT. The Fireworks algorithm (FA) with a low computing time has been successfully applied to ASP, which mainly consists of three steps: initialization, explosion and selection and the final optimization single objective must be consistent with the objective of the subsequence 4 . Also, the precedence graph has been applied, which is a typical expression of assembly interference constraint in product assembly process.…”

“… is the total working time and is the load balancing. To verify the effectiveness of the proposed IDFA algorithm, the computing time and solution quality of IDFA were compared with genetic algorithm (GA) 47 , discrete artificial bee colony algorithm (DABC) 48 , enhanced discrete bee algorithm (EDBA) 49 and assembly subsets prediction method based on precedence graph (ASPM-PG) 4 . The variation probability of GA is 0.1 and crossover probability is 1.…”

“…As the typical manufacturing equipment, industrial robots have been widely used in manufacturing due to their high efficiency and reliability. Robotic assembly has greatly improved the efficiency of product assembly.Assembly-oriented production scheduling is a decision-making process in which limited assembly resources are matched with assembly tasks under certain sequence and time constraints, so as to optimize the assembly scheduling objective 3,4 . Assembly sequence planning (ASP) is needed for production scheduling [5][6][7][8] and it is an NP-hard problem 9,10 .…”

“…There are many studies on ASP. For example, in order to obtain the feasible assembly sequence in the shortest time, Zhang et al 4 proposed an ASP method based on assembly precedence graph, where a simple firework algorithm was used to minimize the number of changing assembly direction and switching assembly tools. Based on ASP, the production tasks can be scheduled.Cycle time is a key factor in robotic assembly line (RAL) task scheduling 11 .…”

A spatial information inference method for programming by demonstration of assembly tasks by integrating visual observation with CAD model

“…Bahubalendruni and Biswal (2016) reviewed the recent researches on assembly sequence generation and found that a robust and efficient methodology is still needed to generate optimal assembly sequences automatically from CAD model without any user intervention. Conventional researches generated the assembly sequence by analyzing the assembly relations, assembly subset, the generation priority and interference matrix obtained from the CAD model (Zhang et al, 2019;Gunji et al, 2018). An optimal assembly sequence planning method was proposed in Wan et al (2018) with the consideration of stability, graspability and assemblability qualities.…”