Model Based Estimation and Tracking Method for Agile Software Project

摘要

In this paper, the authors propose a software cost estimation model for agile software development project which can help estimate concrete development costs for the desired features of a product and tracks the project progress dynamically. In general, cost estimation methods for agile developments used a story point. Because it is relative value, the estimation results tend to be easily fluctuated by the small variation of the base story point. To estimate the effort of a project development, the authors use function points in addition to the story point. The function points are determined based on the desired features of the project. The authors adopt the Kalman filter algorithm for tracking project progress. The daily variation of the function point is observed and inputted to the Kalman Filter for providing concrete estimation and velocity. The authors then validate the better performance of our model by comparing with traditional methods through a case study.