Study On Service Interaction And Application Based On Design Structure Matrix

The current manufacturing industry is characterized by uneven distribution and low utilization of resources and constraints of manufacturing environments. For improving such a situation, accelerating the healthy development and enhancing the international influence of China’s manufacturing industry, Cloud Manufacturing was emerging along with the extension of the kinds of resources, models of services and technologies of Cloud Computing. The interactions among sub-services of a composite service must be studied in order to achieve the goal of providing on-demand high quality services at any time under the circumstance of Cloud Manufacturing, so Design Structure Matrix(DSM) is applied to the process of service composition to plan the interactions among sub-services, to verify the validity of this approach with examples.At the beginning, this paper introduces the theory of Cloud Manufacturing and DSM, and summarizes the steps of constructing and the principles of reconstructing the DSM for a target system according to the existing applications.Then, discussed are clustering and partitioning algorithms that are most commonly used, the clustering of DSM is realized by introducing the correlation cost parameter, and the partitioning of DSM is realized by applying the Path Searching Method. And the use of tearing and banding algorithms are briefly analyzed.Furthermore, with the comparison of the functions and advantages of BPEL, Petri Nets and Coloane, Coloane is selected to transform the BPEL model of service composition into the Petri Nets model. The shop scheduling problem FT46 and a service booking system are taken as examples to verify such a transformation.In the last, the effectiveness of the application of DSM in service interaction aspects is demonstrated through two examples. In the first example, a DSM is applied to service composition process to realize the rearrangement of sub-services, the reduction of feedback associations and the shortening of information transfer chain by using partitioning algorithm. In the second example, a DSM is applied to the design of vehicle engines, and the clustering operation is realized by introducing Bid values and Coordination Cost values in Matlab programs so that the design iterations are reduced, and the overall designing efficiency is improved.