论文部分内容阅读
软件可靠性增长模型(SRGM)是软件可靠性工程中一项重要的研究内容.在可靠性增长模型应用的过程中,常常因为模型假设与实际软件开发和调试过程有差异,导致可靠性预测的准确性不高.至今尚没有一种能适应各种软件开发环境的SRGM.为此,某些国外文献采用遗传(GA)算法,提出了模型组合方法,以期提高SRGM的预测能力.本文采用GM DH神经网络,提出一种非线性的SRGM模型组合方法.通过对比基于GA算法的模型组合方法,实验结果表明,基于GMDH神经网络的组合方法能有效提高模型预测能力.
Software reliability growth model (SRGM) is an important research project in software reliability engineering.During the application of reliability growth model, often because of the difference between the model assumption and the actual software development and debugging process, the reliability prediction Accuracy is not high.So far, there is no SRGM that can adapt to various software development environments.Therefore, some foreign literatures use genetic algorithm (GA) to propose a model combination method in order to improve the prediction ability of SRGM.In this paper, GM DH neural network, a nonlinear SRGM model combination method is proposed.By comparing with the model combination method based on GA algorithm, the experimental results show that the combination method based on GMDH neural network can effectively improve the model predictive ability.