摘要
针对软件老化引起系统资源损耗、系统性能恶化以及软件失效的现象,提出利用强局部加权回归算法进行软件老化趋势的提取方法.该方法利用强局部加权回归算法对系统资源的损耗过程进行提取,得到软件老化的趋势估计,建立软件老化模型.从而可为由于软件老化所引起的软件失效提供前期预测,为系统软件的及时维护提供科学依据.通过对一实际软件系统的实验分析,验证了该方法的可行性和有效性.
Software in operational phase often experiences an aging phenomenon, characterized by either progressive performance degradation or transient failures or both, which is caused by aging-related software faults. In order to deal with software aging, a method based on robust locally weighted regression algorithm was presented to abstract aging trend from system resource usage. With the trend, software fail ure can be predicted and avoided. An experiment for a practical system verifies that the presented method is feasible and effective.
出处
《上海交通大学学报》
EI
CAS
CSCD
北大核心
2006年第11期1951-1954,共4页
Journal of Shanghai Jiaotong University
关键词
软件老化
强局部加权回归
趋势提取
software aging
robust locally weighted regression
trend abstraction
