摘要
日益显现的软件质量问题给高安全、高可靠装备的研制提出严峻挑战,而精确、可信的软件质量预测技术是降低软件测试成本、提升软件质量的有效策略。针对装备大规模、高综合化软件质量与度量之间的强非线性关系,提出一种基于鸽群改进RBF网络的软件质量预测方法。考虑一种软件质量评价体系,以评价软件质量;对于样本提出一种基于主成分分析的数据预处理方法;在所提出的软件质量度量模型中,利用鸽群优化算法改进RBF网络的权值和阈值,以优化训练过程;利用机载软件的质量预测预评价实例,验证了方法的有效性。
The increasingly emerging software quality problems pose serious challenges to the development of high safety and high reliability equipment design.It is noted that accurate and reliable software quality prediction technology is an effective strategy to reduce the software testing cost and improve the software quality.Aiming at analyzing the strong nonlinear relationship between the quality and measurement of large scale and highly integrated software,a software quality prediction method based on a piegon inspired opimization improved RBF network proposed in this paper.Firstly,a software quality evaluation systems is considered to evaluate the software quality.Secondly,a data preprocessing method based on principal component analysis is proposed for samples.Then,in the proposed software quality measurement model,the pigeons swarm optimization algorithm was used to improve the weights and thresholds of the RBF network to optimize the training process.Futhermore,an example of quality prediction and pre evaluation of airborne software is applied to verify the effectiveness of the proposed method.
作者
田川
王闯
刘灿
郭强
TIAN Chuan;WANG Chuang;LIU Can;GUO Qiang(AVICAS Generic Technology Co.,Ltd.,Yangzhou 225000,China)
出处
《航空计算技术》
2021年第5期24-28,共5页
Aeronautical Computing Technique
基金
装备技术基础科研项目资助(202ZX31006)。
