An Improved Comprehensive Evaluation Model of Software Dependability based on Rough Set Theory

摘要

Dependability of software, a major concern inmany computer applications, can be improved throughseveral means. But systematic approaches for its evaluationdo not exist, which is the prerequisite for dependabilitycontrol and improvement. Software dependabilityevaluation is an urgent problem to be solved. There is somesubjectivity about weighting coefficient when applying fuzzycomprehensive evaluation model to evaluate softwaredependability. A new comprehensive evaluation model withcombinational weight based on rough set theory wasproposed in this paper. In this new model, subjective weightwas obtained from expert’s judgment by using thetraditional fuzzy comprehensive evaluation method, and theobjective weight was determined from statistical data byusing the rough set theory. Then, a relatively reasonableweight was obtained by combining the subjective weight andthe objective weight, which effectively avoids subjectivity ofweighing coefficient in fuzzy comprehensive evaluationmodel. In the example, two different dependabilityevaluation results were derived from the traditionalevaluation model and the improved model respectively. Thecomparative result proves the feasibility and validity of thisimproved comprehensive evaluation model.