A Fuzzy Logic Model for Predicting the Development Schedule of Software Projects

The development schedule of software projects is mainly measured in months and it is a necessary and important phase, since the under prediction or over prediction of it at the planning stage can negatively impact budgets. Unfortunately, only 39 percent of software projects finish on time relative to their original plan. According to its development type, a software project can be classified as new, enhanced, or a re-development. In this study, a Fuzzy Logic Model (FLM) for predicting the schedule of new development software projects is proposed. The hypothesis to be tested is that the accuracy of schedule prediction for an FLM is statistically better than the accuracy obtained from a simple linear regression (SLR) model when adjusted function points, obtained from new development software projects, are used as the independent variable. The FLM and SLR models were trained and tested using a sample of new software projects obtained from the International Software Benchmarking Standards Group (ISBSG) Release 11. The accuracy of the FLM was compared against that of the SLR model. The criteria for evaluating the accuracy of these two models were the Absolute Residuals and a Wilcox on statistical test. Results showed that prediction accuracy of an FLM was statistically better than that of an SLR model at the 95% confidence level. We can conclude that an FLM could be applied for predicting the schedule of new development software projects developed on mainframes and coded in third generation programming languages.