A dynamic simulator of software processes to test process assumptions

Validation testing of software processes may provide both qualitative and quantitative suggestions to understand the ways to change the software process to improve its quality, or the ways to achieve speci®c organisation's needs. In many cases, however, such understanding has to be done without a€ecting the actual environment. To this purpose, this paper introduces the use of process simulators for validation testing. To be e€ective, however, simulators must combine the ability to sustain the complexity of the modelling problem with the so-called dynamic estimation capability, that is the capability of representing the dynamics of the simulated process. To achieve such objectives, the introduced simulator is based on the association of three conventional modelling methods (analytical, continuous and discrete-event) into a unique hybrid multi-level new model, called dynamic capability model (DCM). The paper applies DCM in the context of a waterfall-based software process to study the e€ects of three di€erent quality assurance management policies on given process quality attributes, as e€ort, delivery time, productivity, rework percentage, and product quality. The veri®cation of the simulator representativeness is also performed by reproducing empirically known facts in the process behaviour. Ó 2001 Elsevier Science Inc. All rights reserved.