用含仿射操作的遗传算法求解软件可靠性模型参数估计的奇异性问题

IMPLEMENTATION OF GENETIC ALGORITHM INCLUDING AFINE OPERATION IN SINGULARITY OF PARAMETER EVALUATION IN SOFTWARE RELIABILITY

本文将含仿射操作的遗传算法引入到软件可靠性模型的参数估计问题中，与单纯形法比较，它不但较好地解决了以往参数估计中的奇异性，同时也提高了参数估计的精度．特别是按此方法估算出来的参数使得拟合曲线的后沿与软件故障数据曲线的后沿拟合得很好．

In this paper, we apply genetic algorithm inclnding affine operation to solve the singularity of parameter evaluation in software reliablity, and obtain better results compared with other methods. Specially, the post part of fitting carve by using affine genetic algorithm very fits that of software breakdown data carve. It has practical significance in software reliabilty.