A Heuristic Model for Estimating Component-Based Software System Reliability Using Ant Colony Optimization

Most software reliability models for component-based software systems (CBSSs) depend on either the system architecture or component reliability and glue code reliability. It is important to optimize the usage of paths to obtain more accurate estimates of CBSS reliability. By using bio-inspired optimization techniques that are currently being developed, we can find the value of the most-used path for any CBSS. This paper develops a new model for estimating CBSS reliability which accounts for path usage. The proposed model is proved to perform better than other models. This paper proposes a model for estimating CBSS reliability using a bio-inspired algorithm, ant colony optimization (ACO). This model introduces heuristic component dependency graphs (HCDGs), which help to estimate CBSS reliability. This paper also proposes using the reliability of components and assessing their impact on overall system reliability. The results of this paper show that HCDGs give better reliability estimates than other existing methods. A case study of a security management system is presented, along with its results. The proposed approach is fully dependent on CBSS characteristics and provides a good reliability estimation method for CBSSs, taking into consideration both component reliability and CBSS reliability.