Behavioral monitoring of software projects

Abstract

Software projects are known for their high overruns in terms of execution time and budget. The particularities of software projects and also the amount of data produced by the development of the most problematic software projects (the large-scale ones), information regarding work progress mainly, make the monitoring and, consequently, the control of such projects extremely difficult. In this context, the main contribution of this thesis is the modeling of the monitoring process concretized in an integrated monitoring approach based on a monitoring framework for software projects, named the Behavioral Monitoring Framework. This framework enables the automation of the monitoring process, providing also support for the control of such projects. The proposed monitoring framework is centered on the modeling of the behavior towards work of the human resource involved in the project and its main feature refers to the dynamic perspective that it provides over the monitored project through progress forecasting. The first part of this thesis presents the current state of knowledge in the realm of projects monitoring, discussing the main techniques and methodologies employed today for project assessment. The second part of the thesis defines in details the proposed monitoring approach, presenting the Behavioral Monitoring Framework with its concepts, methodologies and modes. The third part is concerned with the primary validation of the proposed Behavioral Monitoring Framework on data that comes from the development of several real-world commercial software projects. The last part of this thesis focuses on the applications that can benefit from the implementation of the proposed monitoring framework.