Átropos: um modelo para predição de riscos em projetos baseado em históricos de contextos

Abstract

The uncertainties in Project Management are constant, making the administration of risk events a strategic requirement. A proper risk management through the history of the projects developed reduces the risk of planning deviations from project time, cost and quality. In this sense, the use of concepts of ubiquitous computing, such as contexts, context histories, and mobile computing can assist in proactive project management. This Thesis proposes a computational model called Átropos that aims to reduce the probability of failure in projects through risk recommendation. The purpose of this study is to show a model to help teams identify and monitor risks at different points in the project lifecycle, as well as a risk categorization, Risk Breakdown Structure (RBS), through an ontology based on activity theory. Where activity theory considers factors related to the context in which the activity is located and activity relationship with other activities. The proposed model follows concepts applied by the Lean method, where the use of historical context of projects is performed a prediction of risks to a project that starts or during the project execution, in this way, the model makes the risk management more agile, since the manager at the beginning of project will be presented to risks that already occurred in similar projects, or that often become problems during project execution. The work presents a differential the use of context histories to recommend risks to projects. The research was conducted through two case studies. The first was carried out with two teams that evaluated the use of the model during project execution. The recommendations were assessed by a project team with 18 professionals, obtaining a result of 72,66% acceptance. In the second case study, the recommendations were compared with executed projects, this research conducted a case study with 17 projects in execution to assess the risk recommendations. Where the Atropos model achieved an accuracy of 82,92% accuracy when compared to projects already being executed. For historical, a database with 153 projects was used by a financial company.