A New Ant Colony Optimization Algorithm for Optimization of Software Reliability

In the last three to four decades several software reliability model has been planned and developed. They played an important role in making the software reliable. Several software reliability metrics have been identified by the researchers and the performance behavior is shown as per the software metrics. The researchers used several approaches for their proposed software reliability model, and all the approaches are doing well in their specific areas. The problems with the most of the models are that they are confined to their specific domain and do not perform better on different context. However, the selection of a good software reliability model can help in this area. To help in this task, it is very important choose an appropriate failure dataset along with selection of good software reliability model. A metaheuristic approach can be fruitful to find the parameter estimation of the chosen model. This pieces of work considers a classical Goel-Okumoto (G-O) model and some failure datasets. A new ant colony optimization method (NACO) has been used to estimate the parameters of the G-O model. The Mean Value Function (MVF) of G-O model is used to minimize the Root Mean Square Error (RMSE) and to find the goodness of fit. Some software metrics have been identified and the performance has been compared with existing metaheuristic approaches. The proposed approach shows outperform behavior than the current metaheuristic approaches. The comparative analysis is done with the help of MATLAB.