| With the development of computer applications, software plays an increasingly important role in the information society, it is a ubiquitous component of our daily life, as well as provides a new way of study and job. Many studies and jobs often depend on the creditable working of software systems. Existing software technology can't meet the rapid development of software quality requirements. Once software failure, it would cause people serious loss of life and property. With the emergence and development of internet technology, Software development, operation and maintenance change from the traditional closed to open and dynamic and changeable. The current study is focused on how to make sure the credibility of software. The credibility notes that the behavior and result of the software are consistent with the expected of user. The credibility of software behaviors is the main performance of the credibility of software.Trust is an essential ingredient of the transaction process. And trust and risk are two closely related factors to make security decisions during transaction process in an uncertain environment that hides risks. The existing software behaviors trust models mostly regard risk as a supplement to trust or neglect risk.In order to evaluate the credibility of software behaviors more reasonably and accurately, a trust model of software behaviors based on check point risk evaluation is presented. Firstly, adopt fuzzy analytic hierarchy process (FAHP) to figure out the weight of every risk factor of check point, meanwhile, assess the modules value according to Markov Chain Usage Model, thus calculating out the risk value of check point and accumulating every suspected risk check point; and then adopt rewarding or punishment mechanism to evaluate a software behaviors trustworthy, which can judge whether software behaviors are credible or not. The simulated experiment shows that this model can distinguish the potential risk effectively in software behaviors, evaluate the risk value trustworthy and provide objective and reliable information to judge whether software behaviors are credible or not. |