An Improved Quality Of Service Evaluation Method In Internetware

With the rapid development of Internet, the environment of computer software becomes open, dynamic and uncontrollable, which is different from the traditional closed, static and controllable. In such an environment, a new type of software named Internetware arises at the historic moment. Along with the increasing maturity of the software development methods and technology based on component, more and more components can be used. Selection of the software component has a great influence on the global software quality of the Internetware, and as an important part of the non-functional properties of component quality of service, its quality directly affect the users on the selection of components. So how to evaluate the the service quality of components and how to select the successful component for users to participate together have become a hot issues.In order to solve this problem, this paper analyzed the characteristics of the Internetware and the basic definitions of software component. This paper put forward a kind of assessment approach with component service quality based on Mahalanobis distance by improving the traditional Technique for Order Preference by Similarity to an Ideal Solution(TOPSIS). In this paper, the main research work includes the following several aspects:Firstly, evaluation of the existing Internetware model and software component model were studied, and the advantages and disadvantages of the present component evaluation model were compared, and the evaluation of the software component model was put forward on this basis.Secondly, in order to reflect subjectivity and fuzziness of the service quality attributes, combination weighting method was proposed based on analytic hierarchy process and entropy coefficient method, which not only reflected the internal connection of the attributes, but also reflected the characteristics of the user’s preferences.Thirdly, to break through the limits of real numbers for the values in an algorithms, the paper extended the method of TOPSIS to interval number, and gave the specific operation steps, which was helpful for the interval type of component of service quality assessment.Fourthly, in order to eliminate the influence by correlation between attribute, traditional method of TOPSIS was improved to TOPSIS based on Mahalanobis distance. Through the analysis of the properties of Mahalanobis distance, we knew that it kept the same results by a nonsingular linear transformation, which was not affected by the influence of dimensional metric and even had an ability to eliminate the correlation of each attribute index.At last, we proved that the TOPSIS component service quality evaluation model based on Mahalanobis distance was feasibility and effectiveness through simulation experiment, and it could select the optimal component for the user to involved in.