广义边际效应在基于覆盖率信息错误定位中的应用

In present, the process of finding bugs in a program is till complex and expensive. Reviwing the technique of being provided, We find that the way to find bugs is always by calculating every statement’s likelihood of having a bug with coverage information. When calculating the likelihood, the weight of increasing or decreasing the number of successful or failed test is the same. In this paper, We propose the general marginal utility model to the existing fault location methods.That means the weight of the second failed test in computing the likelihood of having a bug is larger than the weight of the first in computing the likelihood of having a bug, the weight of the third is larger than the weight of the second and so on . From the studying of the general marginal utility of the number of successful or failed , We find a way to improve the existing fault location method.The experiment results shows that enlarging the weight of failed test is bettter than reducing the weight of failed or/and successful test.