Markov Reliability Model Based on Error Classification

Traditional Markov reliability model used to predict the total number of errors in the software and the failure interval, but did not consider the impact of the severity of the error. However, in some cases, such as software trustworthiness, we must consider not only the total number of software errors and failure interval, but also consider the severity of the error on the impact of the credibility of software. Both in the traditional reliability studies, to consider the impact of the severity of the error, the cost of software development budget or duration of the forecast is very important. Based on the reliability of the traditional Markov model, we introduce the thought of classification of errors into the traditional model, the availability of a new mathematical model of reliability, the reliability of the Markov model is a special case of the model. Using the new model can predict not only the total number of errors in the software and the failure interval, but also can predict the total number of each type of error and failure types, the test indicates a good accuracy, which is the traditional Markov reliability model can not predict.