基于指数平滑模型的大型开源软件开发成本预测分析  

Prediction of the Development Costs of Large-Scale Open Source Software Based on Exponential Smoothing Model

作　　者：许旻鸿[1] 郑巍[1] 

机构地区：[1]广东技术师范学院计算机网络中心,广东广州510635

出　　处：《湘潭大学自然科学学报》2011年第1期123-126,共4页Natural Science Journal of Xiangtan University

基　　金：广东技术师范学院校级科研项目(08KJY06)

摘　　要：Linux和FreeBSD内核进化有较明显的超线性趋势和内在规律.研究其开发成本是管理软件工程的一项重要的基础性工作,找到一种简易又有足够精度的预测方法,科学地预测分析大型开源软件开发成本,有助于大型开源软件发展.以Linux的54个内核和FreeBSD的26个内核版本数据作为时间序列,用指数平滑模型建模并做出预测,将预测结果与最新的Linux和与FreeBSD内核的实际计算结果进行对比,预测结果令人满意,表明该类模型可用于大型开源软件开发成本预测.Linux and the FreeBSD essence evolution have the obvious ultralinear tendency and the inherent laws.It is an important foundational work to study the development costs for managing software projects.In order to help the development of large-scale open source software,we need to find a simple forecast technique of sufficient accuracy to predict the costs scientifically.We take the time series by the Linux 54 essences and the FreeBSD 26 essence edition data,and make the forecast with the index smooth model modelling.Then we canpare the forecasting result with the newest linux and FreeBSD essence＇s actual computation results,and the forecasting result is satisfying.It indicates that this kind of model can be used in large-scale opening the source software development cost forecast.

关 键 词：LINUX FREEBSD 开发成本 指数平滑 Minitab 软件度量 

分 类 号：O431.2[机械工程—光学工程]