Research On Resource Distribution Models And Solving Algorithms For Networking Discrete Manufacturing

In recent years, in order to use all kinds of production factors more effectively and createhigher productivity, the network manufacturing becomes a new kind of enterprise servicemodel. At present, the study of network manufacturing mainly concentrated in the allocationof resources and solving method, but the proposed models are abstract and not reality enough.The solving method can’t solve more complex resource allocation model enough. On thisbasis, the thesis sets up resource allocation models which are much closer to the actualproduction requirements and designs two kinds of improved genetic algorithm according tothe practical application of discrete manufacturing. It improves the efficiency andeffectiveness of solving resource allocation models.First of all, the thesis reviews the research of network manufacturing at home and abroad,and the methods to solve the multi-objective optimization problems. It summarizes the keytechnologies and research fields of network manufacturing, and analyzes the distinction ofdiscrete and continuous manufacturing. So the thesis chooses the discrete manufacturing asthe research target. On this basis, the thesis analyzes the characteristics of discretemanufacturing, and summarizes eight kinds of comprehensive network manufacturing model,and selects the multi-tasking overall optimal allocation model to solve. The models take thenumber of discrete manufacturing enterprises, the number of enterprises’ resources,manufacturing time, manufacturing costs and transportation time into consideration to solvethe multi-objective problem of time and cost’s optimization. Then the thesis studies thecommonly used method of multi-objective optimization, like lexicographic method andNSGA-II. It designs two kinds of improved genetic algorithm, LMGA and SPGA, which areon the basis of Pareto genetic algorithm and lexicographic method. At last, according to aspecific resource allocation case of discrete manufacturing, the thesis solves the model usingSPGA, LMGA and NSGA-II, gives the optimal resource allocation scheme, manufacturingtime and cost for different requirements and analyzes the three results.The experiment shows that the established resource distribution models for networkingmanufacturing considers the key factors of actual production and constraints. The modeldesign is according to the actual production process of networked manufacturing. So it haseffectiveness and practicability. The design of two kinds of algorithm is effective, and theefficiency and effectiveness are better than NSGA-II. The result has a certain guidingsignificance to resource distribution in network discrete manufacturing.