Dynamic Resource-allocation for Software-module Testing

In a previous study, Ohtera and Yamada (1990) formulated and solved two resource allocation problems for software-module testing, considering only the mean number of remaining faults in the software modules. When software-module testing is completed, the actual number of remaining faults may turn out to be much larger than the mean. It is, therefore, important to reduce the variance of the number of remaining faults. In this paper, we study a dynamic resource allocation strategy for software module testing. This strategy takes into account the variations of the number of detected faults during testing, re-estimates the model parameters using all the available fault detection data and dynamically allocates the testing resources to the software modules. The simulation results show that this strategy can effectively reduce the variance of the number of remaining faults.