Resource-Service Sequence Analysis Method Based On Features In Collaborative Tasks

In collaborative tasks,a business process is generally completed by multiple organizations with a cooperative way,and the resource-service management in distributed environment is always a research theme.The way of resource-services provided to business process is “service flow”,which is one of primary methods.However,the selection and utilization of resource-services which are needed by business process are self-completed by organizations,which make the collaboration among resource-services more difficult and gradually become the major constraint on the efficiency of collaborative business process.Therefore,the discovery of resource-service flow has already become an important research theme of collaborative tasks.In order to improve the collaboration efficiency among resource-services and then to improve the global efficiency of business process,we analyze the influence of upstream resource-service on downstream resource-service in Resource-Service Sequence(RSS)deeply based on the features of resource-services and the values of which are mined from business data under the support of workflow technology.And based on this,an approach based on features for the analysis of RSS is proposed.The following studies are carried out respectively:(1)A method for mining frequent resource-service sequence is proposed.Aiming at the temporal relationships among resource-services,and to resolve the problem of producing huge amount of project databases in PrefixSpan algorithm,an improved algorithm of PrefixSpan is studied which is combined with workflow model to mine frequent RSS pattern.Finally,an experiment is designed to test the validity and efficiency of the improved algorithm.(2)An approach for the analysis of key feature sequence of RSS is proposed.For the purpose of obtaining key feature sequence from numerous features,based on the change rate,the influences of upstream resource-service features on downstream resource-service features are studied.As the basis of follow-up processing,the measuring method and solving method of the influence degree are also proposed.Firstly,the clustering method and analysis method for the change rate are studied to obtain the set of key feature sequences,in the analysis method,the continuity of time-series is considered.Then,Based on key feature sequence,the solving method of key resource-service sequence is studied which can improve the selection efficiency of resource-services.Finally,a corresponding prototype system is developed which depended on a cloud manufacturing service platform project to verify the theoretical researches.And the application results of the proposed theoretical methods on the platform which are introduced particularly shows the high performance on resource-service selection.