A dynamic model for evaluating methods to compensate for schedule compression in the system test process

The time available to perform testing is often far less than would be desirable given organizational recommendations. When accelerated schedules are known in advance they can be accommodated, but often scheduling evolves because development tasks take longer than expected. When insufficient time is available to meet organizational standards, new strategies are needed.; This research identifies accepted testing activities and develops two models: an abstract model of the general system testing process, and a dynamic simulation model of an existing organization's system testing process. A series of experiments are performed using the dynamic model to characterize the process performance as the available schedule is reduced. The process is then modified to mitigate the effects of schedule reduction. Experimental results comparing different test strategies are presented and discussed. The results indicate that spending the recommended amount of time on an activity is a better strategy than reducing the amount of time per activity in order to process more items, even though not everything is completed. Doing a good job on a subset of the total items provides more value than doing a poor job on every item.; This research extends previous research in both software system testing and in software process simulation. In the former area, the effects of selected testing activities are studied and analyzed statistically. In the latter area, a systems dynamic model is developed which describes and implements a number of causal relationships and behavioral feedback loops.