Research On Operation Mechanism And Key Techniques Of Cloud Manufacturing System For Small And Medium-sized Enterprise

Manufacturing industry plays an important role in the development of national economy,which is the foundation of a nation,the means of rejuvenated country,and the foundation of a powerful country.With the rapid development of economic globalization and Internet,the trend of diversified consumption and personalized demand promotes the transformation and upgrading of manufacturing industry.Only by continuously improving the comprehensive competitiveness,can enterprises win competitive advantages in the market with shorter delivery time,excellent product quality,lower product price and considerate product service.There are some situations in manufacturing industry,some enterprises are unable to complete orders due to lack of sufficient manufacturing resources,while other resources are been idle due to insufficient manufacturing orders.So resource demanders are eager to obtain external resource in a convenient way,while resource owners want to increase frequency of resources utilization through resource sharing,thereby the relevant enterprises can achieve mutual benefit.The traditional product manufacturing pattern could not satisfy resource allocation requirements throughout the whole product life cycle,so the manufacturing industry needs a new kind of service pattern to achieve reasonable allocation and effective resource usage.The cloud manufacturing,with the goal of resource servitization sharing,has become a feasible way to solve the problem of improving quality,efficiency and upgrading of current manufacturing industry.The emergence of cloud manufacturing provides a new way for manufacturing industry to transform from production type to service type.It integrates some new technologies such as Cloud Computing,Big Data,Internet of things,Intelligent Manufacturing and Knowledge Service.Manufacturing resources and manufacturing capabilities are virtualized as manufacturing services to achieve universal and efficient resource sharing and enterprise collaboration.As a result,the new kind of manufacturing service style is realized,namely "scattered resources are used centrally" and "centralized resources are served scattered".By summarizing the current research status and progress on the basic theory and application of cloud manufacturing at home and abroad,this paper analyzes the shortcomings and limitations in current research on cloud manufacturing systems for small and medium-sized enterprises.Then the operation mechanism and key technologies of cloud manufacturing system in that field are taken as research objects,and the mainresearch contents in the paper are described as follows:Firstly,Manufacturing resources in cloud manufacturing have the characteristics of diversity,wide-area,heterogeneity and servitization encapsulation requirements.Servitization description of resource is an agreement between service provider and requester.In order to improve the accuracy and universality of resource description,this paper proposes a kind of manufacturing resource description method combining the invariance and variability of resources,and the relationship between stability and variability characteristics of resource is well resolved.First,the manufacturing resources are classified and abstract described as manufacturing capabilities,which can be registered and released on the cloud manufacturing service platform through service package.Then the resource operation relation model is established,in which the resource utilization process is described as a set of basic operator with logical relations.As a result,a kind of convenient way for the matching of manufacturing resources and service requirements is achieved.Secondly,according to the phenomenon of resource conflict caused by execution sequence of multi-batch tasks in cloud manufacturing,a kind of dynamic response model of discrete manufacturing events is constructed to solve the conflicts of resource and task scheduling problem.First,task relation matrix(TRM)is used to describe relation among tasks,and then the new task execution sequence has been re-planned through matrix similar transformation.Then,a kind of discrete event dynamic response(DEDR)prediction model is established,and the time dynamic response relation and capital dynamic response relation of manufacturing tasks are derived from them.Finally,a resource scheduling optimization model based on the dynamic response of task startup time is proposed,which can reduce conflict among resources and ensure the orderly schedule of multi-batch tasks and effective use of resources.Thirdly,in order to solve the problem of QoS-based(Quality of Service)resource discovery and service's search space,the constraint propagation relationship among QoS factors and QoS evaluation value are analyzed in the paper.Based on qualitative influencing graph(QIG),a kind of qualitative reasoning method is proposed to predict the trend of QoS.So the service combination scheme with poor QoS value cannot be transmitted to subsequent service procedure.In this way service decomposition andcombination schemes are adjusted according to QoS prediction results.So the service search space can be effectively reduced and the overall operation efficiency of cloud manufacturing system can also be properly improved.Then,according to the problem of QoS fluctuation caused by the unexpected fault of manufacturing service node,a method for determining the weight value of QoS evaluation factors based on objective decision is proposed.When rough set(RS)theory is used to deal with the uncertain and inaccurate problems,it doesn't need any priori information.By classifying the service record data accumulated in the cloud platform,the classification accuracy of QoS attribute is converted into attribute weight value and service similarity criteria are constructed.Accordingly,the candidate service with higher similarity value is selected to replace the failed service node,and the impact degree of fault service node to stability of cloud platform is reduced to some lower extent.Finally,for the functional requirements of cloud manufacturing service platform,the framework structure,database structure and functional flowchart of manufacturing resource allocation management platform prototype system for small and medium-sized enterprise are designed.The main functions of the prototype system are realized through programming,and the reliability of prototype system is proved through experimental results.