Software reliability means the probability of a software to run in a given environment fault-free for a specific time period. The software is an integral part of various economic, manufacturing and military activities. With software application in many security systems, it has now become an important research area. One tool used to assess software engineering technologies is the measurement of software reliability. In the past 40 years, researchers developed many models based on software reliability parameters, such as the probabilistic variance, non- homogeneous Poisson process, and Bayes process, to assess software reliability. While in sometesting situations the models can efficiently calculate software reliability, no model can correctly estimate the software defects in all experimental conditions. In this paper, we discussed more than 100 research papers to analyze the software reliability models using soft computing. From this analysis, we find that the Neural network (NN) as the most suitable technique to find the reliability, and a lot of further research on cuckoo and the fuzzy logic of complex problems is required.

Index Terms— Software Reliability, Soft Computing, NN, Fuzzy Sets, Genetic Algorithm.