Software Reliability Growth Models Selection Based On Multiple Evaluation Rules

Modern people are becoming more and more dependent on computers in their daily lives.Most industries, from automobile, avionics, oil, and telecommunications to banking, stocks, andpharmaceuticals, require computers to function. As the tasks required become more complex, thecomplexity of computer software and hardware has increased dramatically. As a consequence,the possibility of failure increases. As the requirements for and dependence on computersincreases, the possibility of crises caused by computer failures also increases. High reliability isan important attribute for almost any software system.The goal of this dissertation is to survey the use of multiple evaluation rules in the field ofsoftware reliability engineering, especially software reliability growth models selection. With thegoal, this dissertation mainly includes: analysis to current situation that study to softwarereliability models, pointing out problem of software reliability growth models selection, aframework based on multiple evaluation rules for software reliability growth models selectionwas investigated, and three preferences is introduced, with different preference information,different strategy will be used to support the models selection and analysis and recalibrationsoftware reliability models prediction in order to raise accuracy that software reliability growthmodels predict.