Research And Application Of Trustworthy Evaluation In The Choice Of Internetware

With the development of Internet technology, the software scale is growing and the complexity of Software is increasing exponentially at the same time. The cost and the process of the development of software is more difficult to control, the quality of software is more and more unreliable, and the software crisis is following. In order to solve this crisis, M.D. Mcllroy proposed some concepts, such as the software reuse, software component, component factory and so on, to construct the software system in the way of component composition and evolution. However, as an increase in the number of Internetware, how to choose the right components has been become the hot and difficult of current research.From the existing research work, we could know that trustworthy evaluation model is just given rather than concrete implementation plan in some research. Some only considered the reliability of a trusted attributes without considering other aspects. In order to overcome these disadvantages, this paper puts forward a trustworthy evaluation model and the concrete implementation scheme is given. In view of the problem of the prior condition probability of the model, we adapt the method of risk assessment to obtain the reliability of the relevant properties and provide powerful data basis for the trustworthy evaluation model, in order to improve the credibility of evaluation, and make it more scientifically and reasonably guide the choice of Internetware. The main work in this paper is shown below:(1) Aimed at the problems existing in the prior condition probability, I modify it in the method of risk assessment. Firstly I combine MMHC algorithm with Bayesian network to establish a risk assessment mode and do the preliminary processing of risk so I can get the probability of risk source and the conditional probability of risk. Secondly, in order to generate the related incidence matrix, using the algorithm of Bayesian network to calculate the risk probability, and then using the method of hierarchical model to establish the diagram of hierarchy structure. After the related processing of the correlation matrix, I would get the corresponding reliability of all attribute values and the reliability of the whole system module which provides the powerful data basis for credibility evaluation;(2) About how to choose a better Internetware, the model of trustworthy evaluation based on Bayesian algorithm was proposed in this paper. The model analyzes the structure of the component system in the first place, and then divides the system into modules and structure level so we can set up a corresponding trustworthy evaluation system; Secondly, using the method of bottom-up calculation to determine and calculate the credibility evaluation value of each software entities; Finally, the unified evaluation results are obtained through correcting the evaluation indicators, to guide the choice of Internetware.Experiment shows that: The credibility evaluation that is made by the model of trustworthy evaluation based on Bayesian algorithm based on the prior probability which uses risk assessment mode to correct can more objectively and specifically to evaluate Internetware and more scientifically and effectively guide how to choose the Internetware.