Complex Network-based Change Path Optimization For Mechanical Product Design

As customer demand for products continues to increase and companies continue to make internal adjustments,this leads to an increasing number of design changes to mechanical products.Design changes in mechanical products account for a significant part of the final cost of the product and design change activities largely affect the product development cycle and product quality.To minimize the adverse effects of design change propagation,optimizing the paths of design change propagation has become a central task in mechanical product design change.Based on this,this thesis proposes a method to optimize the change path of mechanical product design from several aspects such as component association relationship,change propagation network,single source design change propagation and multi-source design change propagation,etc.The following is the specific work:(1)Study the relationship between mechanical product components and determine the design change dissemination network.Analyse the association relationships of mechanical product parts and classify them into functional,behavioural and structural relationships.The theory of complex networks is studied,and a part of the complex network is analysed to determine the final network type for change propagation.Mapping the association relations of parts to a directional weighted network,and constructing a propagation network model of mechanical product parts.(2)Optimization of design change propagation paths from a single source.Firstly,the theory related to change propagation is introduced,and the reachability matrix is introduced to determine the reachability between two nodes,providing a theoretical basis for the subsequent construction of the propagation model;secondly,a multi-objective optimization model for single-source change propagation is constructed,with development workload,development cost and change risk as the three optimization objectives;finally,considering that the traditional genetic algorithm is prone to the problems of duplicate paths and local optimal solutions,it is improved by a depth-first Finally,considering that the traditional genetic algorithm is prone to duplicate paths and local optimum solutions,the depth-first search algorithm is used to improve it,and a depth-first search genetic algorithm is designed and used to solve the path optimisation model.In order to verify the correctness of the proposed model and the designed algorithm,a simulation example is designed to verify the results and the results are calculated manually.(3)Multi-source design change propagation path preference.Firstly,on the basis of Leader Rank algorithm,the concept of similarity between nodes is introduced and the algorithm of importance assessment of important nodes in SRank network is proposed.After that,the calculation method of component complexity is introduced,and the complexity of each component is calculated according to its parameters in the historical database.Finally,the change propagation probability,part node importance and complexity are combined to evaluate the change propagation intensity and use them as edge weights in the network to build a multi-source change propagation path optimisation model.The enhanced bat algorithm is applied to solve the optimization model of multi-source change propagation paths,and to obtain the optimal solution set for change propagation paths.(4)Using the truck crane trucking part as an example,a parts change propagation network network is established to solve the optimal path under single and multiple source change propagation,and the model and algorithm proposed in the article are validated.