Research On The Risk Evaluation Of Software Projects

With the development of IT industry and the improvement of software scale, the phenomena of delaying delivering end products, overrunning budget and technical defects are becoming more and more serious in software development process. In the process of implementing projects, how to conduct effective evaluation and prevent these risks so as to identify and eliminate the negative factors' effects upon software development belongs to problem of software project risk management.In software development process, risk identification, risk analysis and risk prevention are three key measurements ensuring project success. Risk assessment is an important part of risk management, and now is becoming a management technique in common use in software development and control. This paper studies Boebm's classical risk assessment theory, SEI's TBQ risk assessment method, cost-based risk assessment method and other recent advanced risk assessment methods.During the early phases of the software life cycle or in the research and development process of new-type projects, little or incomplete information of the projects is available. Risks are usually estimated by experts according to experience and prejudice. This paper proposes a new risk assessment model for software projects, based on fuzzy theory, risk weight and experts' trust degree. This model calculates the risk exposure which is always being used to judge the degree of risks. Besides it also measures the combined influence of a certain risk consequence brought by varied risk factors, and the comprehensive influence on the whole consequence by individual risk, which make it convenient for evaluators to measure risk disasters adequately from both macroscopic and microcosmic points of view. Then, the paper takes the face identification and image emulation devices developed by a software company in Qingdao as examples to verify exactness and suitability of the model. The result shows that the assessment model is applicable.