Fault Estimation of SRGM Using Cuckoo Search and Simulated Annealing Techniques

The normal software system reliability estimation ways typically believe assumptions like statistical distributions that are typically unrealistic. The power to predict the amount of faults throughout development part and a correct testing method helps in specifying timely unharnessed of software system and economical management of project resources. The normal approach of fault prediction using software system reliability growth models needs an outsized variety of failures which could not be on the market at the start of the testing. During this research, Goel-Okomoto Model and Yamada S-Shaped models are used and optimization techniques like simulated annealing are applied there on to estimate the parameters of this model. SA relies on organization techniques having its unvarying improvement supported native search. The proposed research uses the more contemporary cuckoo search technique in comparison to older GA as used by the base research. This research will confirm the effectiveness of this technique as compared to other techniques and then the future researchers can apply this technique on other SRGM models like Kapur and Garg model to improve the estimation process. The same technique can also be used in other optimization tasks like job shop scheduling, time table scheduling and software cost estimation technique parameter estimation jobs. The final analysis done from both graphical results and numerical results clearly confirm that the optimization technique Cuckoo search is better than simulated annealing for all data. The SRGM models when compared for results, revealed that the Goel- Okumoto model performed better than Yamada S Shaped model for estimation of failures for given data.