A software test planning simulation

This paper explores a simulation tool that could be used to calculate more accurately the number of work days it would take a software test team to complete a job on a certain amount of stories and compare the results to conventional static methods. Based on historic data, we were able to create variables and parameters for probability distribution functions in the simulation model. At the same time, functions were programmed inside statecharts to predict the amount of work days needed for the test team to complete. Additionally, an implementation of a tester work time off was added into the model in order to provide realistic results. Comparisons were made between the results of the simulation executions and static calculations based on the historic data using Excel spreadsheets for the total number of work days spent by a defined number of testers. The results showed that the difference between the two increased as the number of testers increased. Similarly, when the “out of the office” (time off) option was selected in the simulation, the number of work days was higher. At the same time, it was found that an increase of tester workers decreased the total of workdays and reduced the teams’ time off. We were able to show the potential of a more realistic system that may help with the test planning predictions in a software development environment. The intention of the simulation is to offer a planning and decision support tool for project managers.