Análise de dependência de riscos em gerenciamento colaborativo de riscos

Abstract

In a collaborative risk management environment, which the project stakeholders are involved in risk management activities, the identification and analysis of dependencies often is not a simple process to be performed, although to be necessary for more efficient and concise risk management. This work propose an approach for the identification and analysis of direct and causal dependencies between the risks identified in software projects, being this analysis supported by a probabilistic model of Bayesian networks to assist participants during the process. Thus, collaborative risks discussions of software projects can be conducted through an existing discussion system, the RD System. This research includes the adaptation made in the protocol used by the RD System wherein these modifications by making possible the identification and the analysis of risk dependencies collaboratively. As a way to assist and support the process of analyzing the dependencies once identified, a probabilistic model of Bayesian networks is proposed, enabling the participants simulate different scenarios and discuss starts from the results obtained in a collaborative way with the aim of improve the strategies for the treatment of risks of software projects. In order to validate the proposed approach, we present and discuss a case study involving experts in the area and a practical experiment carried out with students from the computing area. The results show clear evidence of acceptance and applicability of the developed approach for to resolve the identification and analysis of risks dependencies in collaborative environments for risk management of software project.