Research On ADSS Model Composition Process Based On Model Driven And Workflow

With the application from a new generation of software development technology,such as Model-Driven Engineering,artificial intelligence and so on,how to provide an automatic and intelligent model construction environment for agricultural decision support system(ADSS)has become a new research direction.Because currently there lacks a comprehensive research on the domain expert oriented model generation mechanism and environment configuration,and current ADSS modeling researches are mostly too simple and scattered,it creates the situation that engineering and intelligence oriented system is missing.The paper discusses the construction principle and composition mechanisms about the model-driven decision support system(MD-DSS),and focuses on the model composition and solving process with templates,dynamic compilation,and workflow technology on the basis of having known the model expression form.The purpose of the study is to explore the automation process about the ADSS model from the construction to the realization.The innovation of the study could be summarized as follows:1.Research on the automatic calculation of the agricultural decision model based on model-drivenFirst of all,the principle and characteristics of MD-DSS are summarized,and its research and application in agricultural decision making system are summarized.Having known ADSS model expression based on content and form,in computer implementation ways,we discuss the feasibility of the whole process from construction,conversion,composite and computation of the model and related technology.We try to integrate these techniques in order to make the model calculation process to be automated or semi automated.Firstly,the model expression way and their mutual conversion are analyzed,and we mainly summarize the generic of the expression of eXtensible Markup Language(XML)based,highlighting the features and advantages of the Text Template Transformation Toolkit(T4)template in the model expression way.Secondly,the workflow modeling is discussed and the scientific workflow applications in modeling are expounded.Finally,combined with the characteristics of WF workflow,the feasibility of the model composition solution is analyzed.2.Design of computing environment framework based on the workflow of the agricultural decision-making modelOn the basis of the contents of the first part,this paper puts forward a automation strategy based on workflow technology to complete DSS model from the creation to the operation,and implements the corresponding prototype framework named Composite Framework of ADSS Model(CFADSSM).The development process of the framework is designed based on object-oriented thinking and combined with the relevant software development technology.At first,the model's construction is implemented with the third party controls named MathMLControl.Secondly,the model is parsed using the way of recursive descent analysis.Then,the T4 template and the dynamic compiler technology are used to generate workflow model file and reflection mechanism to dynamically generate the corresponding assembly.Finally,with the Event-Condition-Action(ECA)rules revised,the model is viewed as anactivity that is imported into the custom workflow designer,and composition solution of the model is complete by means of embedded Windows Workflow Foundation(WF)engine technology.In this paper,the framework's interface is friendly,the transformation process is transparent,the process of composition is human,and it has certain expansibility.3.Model validation and summarizationThe combined process of the two models of the growth rate model and the apple leaf growth model are tested respectively.Firstly,the CFADSSM framework is used to generate the model activity,and corresponding ECA rules will be constructed.Then the composite work will be done in the workflow designer by using the activity diagram,and finally we enter the relevant parameters to complete the solution of mathematical model.At the end of this paper,three kinds of methods are compared,and the advantage of using workflow is summarized.The article also points out the existing problems and the future development direction.The paper mainly explores and studies in the following aspects:(1)organic fusion of various technologies in the framework of design are used,and model driven engineering ideas are taken on model construction platform and composite environment independence,which can meet the model only need to focus on the business logic without the need to consider the platform the specific implementation details of the requirements;(2)using the idea of workflow model of composite process,this paper proposes a state and data driven method of double factor model compound to make up for the traditional single factor driving range is narrow and multi factor driving method is lack of algorithms or sub models such as lack of hierarchy description.It also satisfies the black box abstraction of model composite process and improves the model compound efficiency.