| "Software Crisis"has plagued software industry for a long time. In order to get rid of it, the method of Project Management has been applied in software projects. But software is abstract and software development process is dynamic, managing software projects is more difficult than other projects. So we must put more attention on measurable management work, it can control and predict software development process more effectively. It also can provide decision-making for balancing quality, cost and schedule in software projects.Because of this reason, it is important to implement measurable management work based on software reliability in software projects. Software reliability is the most important quality characteristic and easy to be measured. Software is dynamic, so it suit to evaluate and predict software development process. Software reliability is related with quality, cost and schedule in software projects. Software reliability prediction model is the base to evaluate and predict software reliability.The content of software project management is mentioned in this paper. Then the effect of software reliability is discussed in software project management. Evaluating and predicting software reliability can offer decision-making for balancing quality, cost and schedule in software projects, so if reliability prediction precision can be improved, it will make software project management more effectively.Reliability prediction precision is not better in the traditional software reliability prediction models .In order to resolve this problem, semi-analytical prediction method is established in this paper. The new method combines Artificial Neural Network (ANN) and reliability growth model. It is proved that the semi-analytical prediction method makes the model has a better prediction precision, a better generalization ability and a lower dependence on the number of samples by using Matlab tool.In the end, the author indicates the shortage in the research and prospects the next work. |
PDF Full Text Request