Research And Development Of System In Risk Control Based On UML

With the increasing requirements in software, there are more and more problems in software development which lead to project delay, high expense and low quality called software crisis. In the process of software development, risk is inevitable, so it how to define risk, evaluate and measure risk and make out compensative strategies in order to decrease risks to an acceptable level is an essential problem to solve.Firstly, the all kinds of risk management models are introduced, meantime, a new risk management model is proposesd which comprises risk identity, risk analyze, risk plan, risk track, risk control and risk learn. The model combinates the merit in CRM of SEI and MSF of Microsoft and emulates their shortcomings.The important point to research is the process of risk control; risk in the development is deseased from different views to achieve objects of development. The Full Function Point based on UML is proposed to measure size of software exactly; requirement formalization by B method based on UML is proposed to emulate the incomplete, inconsistent and ambiguous in software requirements because problems in requirements are fatal. A method is proposed which reconciles the relation of schedule, cost and quality to achieve the final object in software engineering that software production is developed in time and cost scheduled based on research on Fuzzy estimation, Bayesians networks, Colored Petri Networks and PERT(program estimation and review technology).Finally, a system is developed to measure software size, sort in risks of projection and realize functions related.