Ranking of components for reliability estimation of CBSS using fuzzy TOPSIS

Abstract

Component-based Software Systems are comprised of reusable components. These components are independent and replaceable entities. The overall reliability of a Component-based application depends upon the reliability of these components and the reliability of their glue code. Every component does not contribute equally to the reliability of a component-based application. Each component has a different usage frequency; hence the reliability of a Component-based application depends upon the usage frequency of each component. This usage frequency is the number of times a component is executed during a typical operation. However, a component may be ranked on the basis of its reliability impact on overall application reliability. Ranking components is a multi-criteria decision problem. In this paper we ranked the components by using a Fuzzy technique for order preference by similarity to ideal solution (TOPSIS). The weight criteria and ratings of alternatives are assessed by linguistic variables expressed in terms of triangular fuzzy numbers. By this proposed technique we may rank the components according to the effects of their reliability on overall CBSS application reliability.