Computer Engineering and Applications ›› 2008, Vol. 44 ›› Issue (11): 47-49.

• 理论研究 • Previous Articles     Next Articles

Estimating parameters of software reliability models using PSO

ZHANG Ke-han1,LI Ai-guo2,SONG Bao-wei1   

  1. 1.Department of Marine,Northwestern Polytenical University,Xi’an 710072,China
    2.Department of Computer,Northwestern Polytenical University,Xi’an 710072,China
  • Received:2007-10-09 Revised:2008-01-02 Online:2008-04-11 Published:2008-04-11
  • Contact: ZHANG Ke-han

基于PSO的软件可靠性模型参数估计方法

张克涵1,李爱国2,宋保维1   

  1. 1.西北工业大学 航海学院,西安 710072
    2.西北工业大学 计算机学院,西安 710072
  • 通讯作者: 张克涵

Abstract: It is an important research field that models software reliability.Presented software reliability models are almost nonlinear function models,so it is difficult to estimate their parameters.Particle Swarm Optimizers are valuable stochastic optimization methods for various solving nonlinear optimization problems.A method of estimating parameters of software reliability models based on Particle Swarm Optimization is proposed in this paper.The key of this method is to construct fitness function for estimated software reliability models.Parameters of exponential software reliability growth model and logarithmic Poisson execution model of five software systems are estimated using proposed method respectively.The experimental results show that estimation precise of proposed method is high.

Key words: software reliability model, Particle Swarm Optimization(PSO), parameter estimation

摘要: 软件可靠性建模是一个重要的研究领域,现有的软件可靠性模型基本上是非线性函数模型,估计这些模型的参数比较困难。粒子群优化是一类适合求解非线性优化问题的随机优化方法,提出一种基于粒子群优化的软件可靠性模型估计参数方法,该方法的关键是构造合适的适应函数。用该方法分别估计了5个实际软件系统的指数软件可靠性模型以及对数泊松执行时间模型,实验结果表明:该方法参数估计的精度高,对模型的适应性强。

关键词: 软件可靠性模型, 粒子群算法, 参数估计