A Study on Probabilistic Schedule Estimation and Refinement of the Small and Medium-Sized Software Development Projects

Abstract

This paper proposes a research framework for objective and probabilistic project schedule estimation and refinement targeted at a small and medium-sized project, as well as, verification of the research framework. Graphic User Interface (GUI) and design documents, which are composed of a class diagram, sequence diagram, and normalized database, are created to count the accurate number of Function Point (FP). The counted FP, which is the main input for calculating Adjusted Function Point (AFP), is then used to estimate an accurate project size, recalculate human resource efforts, and yield single point-based project schedule estimation. According to programmer's subjective judgment, application development is supposed to be completed within 10 business days. Yet, objective 3-point schedule estimation shows that the success rate of completion within 10 business days is infeasible. Additionally, single point estimation shows that at least 1.313 months (approximately 30 business days) is required to complete application development if more than 80% success rate is expected. The effectiveness of the research framework is successfully demonstrated though development of the actual Android OS-based mobile application (EzMaintenance.apk).