Design And Implementation Of Workflow Intelligent Process Engine For Customer Service System

The project is derived from the actual project of Beijing Hao Wireless Technology Development Co.,Ltd.With the continuous improvement of people's living standards and the popularity of smart phones,the use of mobile terminals has occupied all aspects of people's food,clothing and housing.More and more users use taxi software for daily travel activities.To be more Convenient and efficient processing of customer feedback information,the project team decided to develop a configurable intelligent process engine to automatically obtain and process user feedback data according to a predefined process model,which is convenient for improving customer service staffs office efficiency.R&D and.maintenance personnel manage and monitor customer service data.To achieve a configurable process engine,the theoretical foundation is to select the appropriate business process modeling standards.Therefore,this paper first studies the business process management technology and decides to choose the business process modeling mark bpmn2.0 as the modeling standard of the process engine.Then through research and comparison The popular business process engine jbpm and activiti comprehensively compare their respective advantages and disadvantages,and combine with the actual business situation of the company,and finally decided to independently develop a process engine suitable for the internal business of Didi.The biggest feature of the process engine is configurability,using web version dynamic form technology and a graphical model designer to configure models that conform to actual business processes.The business operator can quickly and easily configure the business requirements into a business process model through the model designer,through the dynamic form module.The custom form data in the business process model is converted into a data stream recognizable by the engine,which is processed by the engine and then rendered as a dynamic form.Through the introduction of design goals,the paper defines the functional and non-functional requirements of the engine.The engine is divided into five modules:scene management module,process management module,form management module,tool management module and engine scheduling module.Based on this,the overall architecture,layered structure,and overall process of the engine Database tables and interfaces have been planned and designed.Then,the engine scheduling module responsible for the author is designed in detail,and the scheduling module of the engine is subdivided into a main scheduler module,a node scheduler module,an action scheduler module and a data scheduler module according to the life cycle of the data flow.By targeting The sub-module class relationship and data flow relationship are designed in detail,and the function of returning to the caller system in the form of sdk after the operation of the engine scheduling module by the configured business flow chart is realized.Finally,the engine is tested for functional and performance.The functional test mainly tests the flow of data between multiple processes,multiple forms,and multiple nodes in the scenario under the "additional surcharge" scenario.The performance test is mainly to obtain the performance monitoring data of the key interface of the engine by the Drip "Odin" monitoring system.The test results show that the engine not only achieves the expected function,but also has good response speed,and achieves the performance index in the non-functional demand analysis in the case of high concurrency.After the company's customer service efficiency department evaluates feedback,the engine covers more than 70%of the work order system after the online service,and directly or indirectly improves the customer service efficiency by about 73%.