Improved Cosine Similarity-based Artificial Bee Colony Optimization scheme for reactive and dynamic service composition

Abstract

Reliable and dynamic composition of web services is considered as essential for ensuring continuous services to the users since they are responsible for integrating varsity of applications in spite of their independency. The significant development in web services domain in the last two decades enable the option of devising novel service composition and service selection schemes for optimal performance and success rate of dynamic web service composition. Majority of the research works proposed for dynamic composition of web services was confirmed to be formulated using the characteristics of Quality of Service (QoS) or Transactional features of workflow. Improved Cosine Similarity-based Artificial Bee Colony Optimization Scheme for Web Service Composition (ICS-ABCO-WSC) is proposed for integrating the characteristics of QoS and transactional features for determining optimal candidate service solution from the workflow modeled graph generated during reactive service composition. ICS-ABCO-WSC is proved to enhance the rate of exploitation and exploration by incorporating opposition learning in employee bee phase and, combinatorial search strategical equations and enhance rate factor in the employee and onlooker bee phase respectively. The success rate and optimality index derived using experimental investigation of ICS-ABCO scheme is proved to be 26% and 32% excellent to compared baseline graph-modeled web service composition techniques. This improvement is realized in ICS-ABCO-WSC due to its potential of enhancing the precision and acceleration rate of converging solution achieved in Artificial Bee Colony Optimization technique.