Research On Workflow Enginebased On Dynamic Demand

With the development of information technology in recent years,e-government is applied to governments and enterprises more and more widely.Workflow technique whose intention is to improve the running speed of e-government's website is the essential part of e-government and it also develops rapidly to adapt the evolution of this society.However,some typical and open-source workflow engines,such as JBPM and Activiti,have the following shortcomings.On the one hand,they cannot meet the need of rolling back in domestic business process.On the other hand,when those workflow engines run on server,it will cost too much memory of this server to adapt imbalanced development of domestic computer equipment.In order to meet the dynamic requirement of domestic workflow engine when processing some affairs,we propose a new workflow engine which can adapt to domestic circumstance better.The main work in this paper is as follows:(1)Most work flow engines have a low separation degree distinguishing business data from process data and cannot complete the whole workflow process trans-departmentally.In order to overcome those shortcomings,our engine combines service with service layer and engine kernel layer in architecture design.Besides,we also detailedly distinguish different layers,such as peripheral layer,external interface layer and interactive proxy layer,to separate business data and process data more precisely.This method can enable our engine to meet the need of workflow management system which may change dynamically.Furthermore,adding trans-department processing layer in model design enables different departments to exchange data and improve business flow rate and the degree of information integration and sharing in actual workflow process.(2)In order to guarantee that work missions can be complete efficiently,our engine are designed and programed with the following aspects.Firstly,automatic task assignment strategy based on multiple principles enables our engine to assign tasks to the most suitable staff by calculating information about staff role,load predicting and user range.Secondly,task rollback scheme based on multiple staves can meet the need of rolling back in actual work by selecting the former staff more precisely using multiple selection rules.(3)In actual workflow process,it often has to grant different authority to different user.In order to grant data more safely,our engine is designed and programmed with a new authority control model based on BRAC and authorization towards granularity,and the data is displayed in a kind of dynamic form based on table parsing,which guarantee users to legally visit data in the workflow process according to their own authority.(4)Our workflow engine has been put into use.The practical data shows that our workflow engine has the following advantages.Firstly,in architecture designing,our engine has more precise and more clear layers,which makes our engine handle affair better,affairs including dynamic requirements and trans-department workflow.Secondly,performance data shows that our engine cost less memory,processes business faster and is more suitable to the circumstance where computer equipment development is imbalanced in our county.