Research On Trustworthiness Measurement Models Based On Software Component

Today's world is an era of wide application of computers and rapid development of networks.Software system has become a key part of the world.It plays an important role in military,economic,political,cultural and people's daily life.However,the system is becoming more and more complex,and the trustworthiness of software systems has attracted more and more attention.The trustworthiness of software is an important index reflecting software quality and product competitiveness,and it is also one of the core scientific issues in the basic research of software trustworthiness.With the rapid development of software component technology,component-based development method has become the mainstream method of software development.The use of component technology is an inevitable experience in the development of software engineering and industrialization.Because of the increasing size,complexity and variety of software structure,it is difficult to measure the trustworthiness of large-scale systems directly.Measuring the trustworthiness of the whole component software system through the trustworthy attributes of components and the relationship between components has become a research hot topic.At present,the research work of component is more focused on software component development and reuse technology,and the research and practice of component trustworthiness is less.In the process of component development,the trustworthy attributes of components have been obtained.How to use the credible attributes of components to get the trustworthiness of component software system has become an urgent problem to be solved.Because trustworthiness is the synthesis of multi-dimensional attributes,the calculation of different trustworthy attributes in component relationships is different.Therefore,this thesis adopts the point,line and surface research methods,from component measurement elements to the trustworthy attributes of component,then to the trustworthy attributes and the trustworthiness of component software system.The main contents of this thesis are as follows:(1)This thesis a quantitative method of component performance specification.The refinement relationship between performance elements is introduced,and the quantitative method based on component performance specification is further studied.Quantitative method of performance transforms the satisfaction problem of performance from qualitative to quantitative.(2)This thesis presents a method to determine the weight of trustworthy attributes based on positive and reciprocal matrix aggregation.The positive and reciprocal matrices are constructed by experts' evaluation of the relationship between different attributes.The positive and reciprocal matrices of aggregation are obtained by the aggregation method based on Euler distance,and then the weight of attributes is obtained by calculating the positive and reciprocal matrices of aggregation.The weight is the comprehensive evaluation of multiple experts.This method keeps good properties and reflects the rationality of the method.The effectiveness of this method is illustrated by a case study.(3)This thesis presents a trustworthiness update model based on users' feedback.In the process of component usage,component trustworthiness measurement is updated by user feedback trustworthiness and component trustworthiness.The model can reflect the usage of components.The model can well reflect the changes of user usage number and trustworthiness in use.(4)This thesis presents trustworthiness measurement models of component-based software system.The trustworthy attributes of component software system are calculated by trustworthy attributes of component and component relationship,and then the trustworthiness of component software system is further calculated.Through progressive research,a series of measurement models are obtained.According to the measurement,the weakest trustworthy attribute can be pointed out,and the component that need to be improved for this trustworthy attribute can be further analyzed.The model has good applicability.