Quantitative Analysis of Value-Based Software Processes Using Decision-Based Stochastic Object Petri-Nets

Abstract

The value-based software process (VBSP) is gaining more and more attention. However, the quantitative analysis techniques for VBSPs could not closely follow up the fast developing paces of the modeling techniques. In this paper, we proposes a decision-based stochastic extension of object petri nets (OPN) to resolve the issues. OPNs are well suited for modeling VBSPs and stochastic object petri nets (SOPN) combine the benefits of OPNs and the stochastic theory. The decision-based stochastic object petri net (DB-SOPN) model is economics driven and links value creation with decision making, multi-stakeholder satisfying, and risk management. It includes two levels: the high level models the guideline of the software process life cycle; and the low-level represents the different stakeholder's perspectives of the process. Some activities of a process have candidate policies that will produce different value reward. Our model simulates the entire software process, and compares various combinations of candidate policies to make the value reward of the process maximum.