Achieving on-time delivery: A two-stage probabilistic scheduling strategy for software projects

Abstract

Due to the uncertainty of software processes, statistic based schedule estimation and stochastic project scheduling both play significant roles in software project management. However, most current work investigates them independently without an integrated process to achieve on-time delivery for software development organisations. For such an issue, this paper proposes a two-stage probabilistic scheduling strategy which aims to decrease schedule overruns. Specifically, a probability based temporal consistency model is employed at the first pre-scheduling stage to support a negotiation between customers and project managers for setting balanced deadlines of individual software processes. At the second scheduling stage, an innovative genetic algorithm based scheduling strategy is proposed to minimise the overall completion time of multiple software processes with individual deadlines. The effectiveness of our strategy in achieving on-time delivery is verified with large scale simulation experiments. © Springer-Verlag Berlin Heidelberg 2009.