A method for stochastic modeling the software development process in constrained resource environments

Process improvement requires a formal process definition and associated measures of performance. Processes can then be diagnosed in order to formulate hypotheses on where changes should be made. Subsequently, new processes can be conceived, and developers can check whether the desired improvements have actually been achieved. Simulation is a powerful low-cost tool for diagnosis and test of several improvement alternatives, prior to field tests. This paper presents a UML-based method to obtain the probability distribution of the execution time of a large variety of business processes, including software development. Such a method, which is based on Monte Carlo simulation, allows for the identification of factors that most strongly influence process execution time, favoring changes that increase process efficiency with considerable impact on the deployment of business tactics and strategies.