Research On Assembly Sequence Planning Based On Discrete Firefly Algorithm

Product assembly is the last order of the manufacturing process, and the cost ofassembly account for a large proportion of the manufacturing cost, even more important isthat the quality of product assembly plays an important role in the performance of theproduct. Furthermore, the cost and quality of assembly depend on the assembly sequenceplanning of product. Therefore, the assembly planning occupies an important position in theproduct design and development. On the one hand, it determines the assembly order ofproduct. On the other hand, it provides a basis for assembly equipment and assembly fixtureselection. It is of importance that the assembly sequence planning method based on fireflyalgorithm is proposed in this thesis. The evaluation indicator of process precedencerelations is also proposed. What’s more, the traditional evaluation indicators areappropriately improved. The experiments in this thesis show that the assembly sequenceplanning method based on discrete firefly algorithm is feasible and effective, the evaluationfunction in this thesis is reasonable and reliable. The main work of this thesis is as follows:(1) The firefly algorithm is applied to the assembly sequence planning in this thesis:A method of encoding for firefly position oriented to assembly sequence planning problemis proposed. Some operations in the position update formula of the firefly algorithm areredefined, by which the discrete firefly algorithm for assembly sequence planning problemis proposed. What is more, a method of setting parameters of discrete firefly algorithm isalso studied and discussed in this thesis, and the general steps and method of settingparameters are proposed. The effects of the capacity of firefly population and the variationof parameters to the performance of algorithm are also deeply studied via an assembly case.(2) To improve the optimization ability of discrete firefly algorithm in a great extent,the iteration rule of the firefly algorithm is improved, and the improved discrete fireflyalgorithm oriented to assembly sequence planning problem is proposed. It is verified byexperiments that the improved discrete firefly algorithm is superior to discrete fireflyalgorithm and genetic algorithm.(3) An evaluation indicator of process precedence relations is proposed in this thesis.The traditional method of quantizing the stability of assembly and solving scheme of the frequency of assembly tool changes are also improved. The experiments show that theevaluation indicators in this thesis are reasonable and reliable.