In software process improvement, accumulating and analyzing the
historical data from past projects are essential work. However, setting up the
systematic and logical measurement and analysis program is very difficult.
Many mature organizations have their own measurement program for the process
improvement. However, most of them are based on the statistical metricsdriven
approach that consequently limits logical reasoning on the detailed
analysis on the process. In this paper, we propose a process analysis approach
called MPAF(Model-based Process Analysis Framework), based on formal
process modeling. In MPAF, the corresponding formal process instance model
is recovered through data gathering from a project execution. Various formal
analysis can be performed on the recovered and reconstructed process instance
model for diagnosing the vitality of the project. We also performed experimental
case study by applying MPAF to real world industry projects.