Research On Methods And Models Of Software Project Risk Management

Software project risk management which is one of key factors relating to software project success is a very important task during software development process. At present, there is basically mature in the research on the fundament theory of software project risk management, but the effective, practical and systematic methodology still lacks. With the rapid increase of software requirement and complexity, it is an urgent and actual problem to research on software project risk management method combining with the current status of software industry and the characteristics of software project.The paper has worked on a series of questions on risk assessment, risk response planning and risk control during software project risk management. The achievements and main works of the dissertation are as follows:Firstly, based on the analysis of defects of the existing software project risk assessment methods, extension assessment method for software project risk is put forward. It contains project comprehensive risk extension assessment and risk event extension assessment. Numerical example demonstrates the practicability and effectiveness of the extension method. The method can reveal different levels of risk as well as the tendency of risk transform, which help project risk managers early detect and prevent risk and make out rational software project risk response planning.Secondly, due to software projects having high complexity and uncertainty, the information about quantitative assessment criterion given by experts has ambiguity. Extraction the experts'information by accurate values will results in loss of information. So group fuzzy analytical hierarchy process based on triangular fuzzy number complementary judgment matrices and software project risk event prioritization based on group fuzzy analytical hierarchy process are proposed. Numerical example analysis shows the maneuverability and validity of the method.Thirdly, based on the analysis of software project risk response actions generation mechanism, according to thought and method of extension planning, the extension model of software project risk response is made and risk response actions are found by extension transformation theory. Thus extension planning method for making risk response actions describing by formal language is given. Fourthly, for optimal selecting problem of software project risk response actions, an interval model is built and its algorithm is also given. The model denotes the parameters by interval numbers, so it is practical. The model minimizes the target of risk response cost and risk level from the whole project perspective. This takes account into not only the cost quantity of risk management but also the decision maker's risk preference. The model provides an effective quantitative analysis tool to help risk managers to make scientific decision.Finally, the current critical chain management method based on probability theory requires giving the probaility distribution of activity duration. It is difficult for software project which lacks of history statistical data and has unique. Aiming at the problem, a fuzzy critcal chain management for software project schedule risk integrateed with fuzzy theory and risk quantitative methods is proposed. A case demonstrates the method feasible and effective. From the overall the project perspective, the method based on project work breakdown structure analyzes schedule risk for project activity and set buffers size accoroding the schedule risk, managing software project schedule risk through monitoring buffers. The method evaluates activity duration by fuzzy language, fully taking account into the ambiguity of human brain mind and software proejct uncertainty and complexity, thus ehancing its practicability.