Genetic Algorithm for Assembly Sequences Planning Based on Heuristic Assembly Knowledge

To improve the efficiency of Assembly Sequences Planning (ASP), a new approach based on heuristic assembly knowledge and genetic algorithm was proposed. First, Connection Graph of Assembly (CGA) was introduced, and then, assembly knowledge was described in the form of Assembly Rings, on that basis, the assembly connection graph model containing Assembly Rings was defined, and the formation of initial population algorithm was given. In addition, a function was designed to measure the feasible assembly and then the genetic algorithm fitness function was given. Finally, an example was shown to illustrate the effectiveness of the algorithm.

Petri Net Models for Assembly Sequences Planning

Assembly sequences can be represented by a Petri net(PN) which characterizes dynamic system changes and provides a tool for obtaining optimal assembly sequences. In this study some assembly operation constraints are considered in order to obtain more practical sequences which are conformed to real situations. In order to enhancing the efficiency of the assembly sequence planning, knowledge-based Petri net, combining an usual Petri net with expert’s knowledge and experiences, is proposed to construct the assembly model. With the complexity of the product, the product’s assembly model size will be too large to analysis. So the basic subnets are used to reduce the large PN. And the reduced version can be used for the analysis of the original PN. To verify the validity and efficiency of the approach, a variety of assemblies including some complicated products from industry are tested, and the corresponding results are also presented.