Visual Hierarchical Use Case Modeling And Evolution Method Based On Graph Transformation

UML (Unified Modeling Language) is widely used in the softwaredevelopment field as a standard visual modeling language, and is one ofsignificant study subject of visual modeling and model evolution in recentyears. Extensively used UML use case model is difficult to evolveprimarily because of the lack of a strict and consistent syntax andsemantics specification on use case model, and this heavily affects thequality of use case model and the modeling efficiency. With the enlargingof software scale, the increasing of software complexity and theaccelerating of software change frequency, UML use case model becomesmore and more difficult to adapt to these situations.The dissertation primarily studies the use case modeling and evolvingbased on graph transformation theory. First the research and applicationbackgrounds are investigated, which include visual use case modeling,software evolution and graph transformation theory. Then a visualhierarchical use case modeling and evolution method based on graphtransformation is proposed, and the related models, techniques, processesand tools are systematically and deeply studied, and moreover, theapplication of this method into access control system shows its validity.A hierarchical use case model based on reuse is proposed, and theextended Backus_Naur Form about the text description of the hierarchicaluse case model is presented. The dissertation defines the abstract syntaxand the corresponding concrete syntax through applying the attributedtyped graph and the graph transformation theory in order to establish thevisual graphical description of the visual hierarchical use case model, andbuilds the algorithms which are utilized to generate the text descriptionfrom the graphical description. Moreover, the denotation semantics of thetext description of the hierarchical use case model is studied integratingwith the event sequence expression.In order to improve the expression capability and the performance ofthe graph transformation in applying to software evolution, a softwareevolution operation description language SEOL (Software EvolutionOperation Language) is proposed, and the structured operational semantics on SEOL is studied, and then the semantic equivalence analysismethod about software evolution operation is described. To the SEOLexpression, the dissertation builds the corresponding processingmechanisms, which include the basic principles and the processing flow,and discusses the graph transformation description in SEOL.According to the characteristics of the hierarchical use case modeldescription, the dissertation proposes an evolutionary use case modelingprocess model, which divides the use case modeling and evolving processinto four phases, namely initial modeling, extending, analyzing, andrefactoring, and fully discusses each phase's activities making use ofattributed typed graph and SEOL. The extending operation rules and theircorresponding undo rules are defined in extending phase. In analyzingphase, the calculation methods and the heuristic rules are proposed for themeasuring of use case size, use case model structure complexity, and usecase model reusability. In refactoring phase, the dissertation defines therefactoring description template, and systematically discusses thehierarchical use model re factoring.At the same time, the dissertation also studies the automationproblems for the process of the hierarchical use case modeling andevolving. Through investigating partial order planning and fuzzyclustering techniques, the designing and the optimizing to use case modelevolution operations are studied. The prototype tools of visual use casemodeling and evolution are discussed according to the hierarchical usecase model description syntax and SEOL.In conclusion, the research works in this dissertation are significant inimproving the use case model quality and enlarging the use case modelapplication scope, and provide the academic reference for visualmodeling and software evolution, which can be widely used in the nextgeneration software development method.Finally, the research works are summarized, and some future researchdirections are also presented.