Research on assembly sequence planning method based on artificial neural networks

Abstract

Assembly sequence planning is acknowledged as a pivotal aspect of the assembly process, playing a vital role in the realization of production automation and the assembly and disassembly of equipment. It is a current research focal point within both national and international realms of Computer Integrated Manufacturing Systems (CIMS) and equipment maintenance domains. This paper introduces and executes a neural network-based methodology for automated assembly sequence planning. The proposed method initially employs assembly interconnection matrices, assembly graphs, and penalty matrices to delineate the preferential relationships and assembly costs among components. Thereafter, it utilizes a Backpropagation (BP) neural network to resolve the optimal product assembly sequence adhering to these constraints. The efficacy of this method provides a substantive contribution to the automation of assembly sequences and is potentially beneficial for optimizing complex assembly tasks in industrial applications. Future work may extend towards refining this approach and testing its adaptability across various manufacturing scenarios.