Research On Dealing With Uncertainties In Enterprise Information System Engineering

In software engineering, various uncertainties from human and environment factors frequently result in long run-time, high risk and consequently often failure, poor quality or unsatisfied implementation of IT projects. This thesis mainly researches on the uncertainties of software development process and uncertain data being integratedly processed in enterprise information systems.The uncertainty on information system development process is the main focus of researches. In software enginnering, software requirements are the most difficult and most influential issues though they are the foundation of further software development activities. The communication and cognition issues from stakeholders, the contextual or environmental issues of operations, the dynamic or development of requirements bring uncertainties for requirements engineering (RE) and consequently uncertainties in software development process. In enterprise information system engineering, the requirements, especially functional requirements are derived or extracted from the related business processes. The alignment of information system requirements with business processes and further connection with deliberate business strategy are important. Upon studying current techniques in requirements engineering and the environments of new generation enterprises, a domain-knowledge based essential requirements approach is proposed. Such approaches reuse the contextual domain-knowledge or ontology, and ensure the integration of strategic business processes with software requirements. Based on the proposed domain-knowledge based RE framework, an information system engineering project is presented as a case study, in which the ARIS (ARchitecture of Integrated Information Systems) process databases are used for requirements elicitation and analysis.Regarding the uncertainty of data, the application of integrating fuzzy data with conventional information systems is studied. The fuzzy information is typically caused by fuzzy concepts and linguistic communication from daily business operations. The application of fuzzy set theory - in connection with expert knowledge - in dealing with uncertain data and complex issues of real-world enterprise operations, the principle and basic process for fuzzy decision making and fuzzy information processing are discussed.Researches in this thesis cover requirements and analysis phases of software life cycle. Based on actual information system projects in the company, in-deep studies on requirements elicitation and analysis as well as synthetic processing techniques of fuzzy information towards decision making are presented. The proposed domain-knowledge based, reuse oriented RE approach and fuzzy sets based integrated application technique on dealing with uncertainties in enterprise information system engineering are illustrated to be useful in practice.