摘要
针对传统软件项目预警的分析的缺点,引入蚁群神经网络算法来进行软件预警评估,从进度、成本、质量3个角度建立预警评估指标。通过蚁群神经网络算法对项目流程进行中的各个环节点进行分析,设定预警信号点辅助决策者对整个项目进行控制。以某公司2015年间所有开发项目为数据进行仿真,得出的预警评估数据要更贴近于软件开发项目的实际风险情况。通过预警评估定量分析后得到的数据,项目经理可以分析诱发软件项目危机的原因,对软件项目进行改进。
In view of the shortcomings of traditional software project warning,this paper introduces the ant colony neural network algorithm to carry out software early warning evaluation,establishing early warning evaluation index from the schedule,cost,quality three angles.Analyzing the project process of each loop node through the ant colony neural network algorithm,it sets the early warning signal point to control the whole project.With a company for 2015 years of all development projects for the simulation of the data,this paper draws the early warning assessment data to be much closer to the actual risk of software development projects.Through the data obtained after the quantitative analysis of the early warning,the project manager can analyze the causes of the project of the software project,and improve the software project.
出处
《微型电脑应用》
2016年第7期74-77,80,共5页
Microcomputer Applications
关键词
软件过程
预警评估
项目经理
神经网络
Software Process
Early Warning Evaluation
Project Manager
Neural Network
