Integrated Ant Colony and Artificial Bee Colony Optimization Meta Heuristic Mechanism for Quality of Service Based Web Service Composition

摘要

The explosion of cloud computing has demanded the service consumers to rely on the multiple web services that are available on the internet that come up with the same functionality. Service Driven Structures (SDSs) are revealing their efficacy in a progressive path exhibiting their vital role in distributed environments like intelligent research domain, health domain, defense domain and aerospace domain by the use of reliable services. Web services are application modules required for the seamless development of web applications. Majority of the works contributed for reactive web service composition depends on transactional characteristics of workflow or QoS constraints of workflow. The drawbacks faced in case of Meta-heuristic algorithms namely Ant Colony Optimization algorithm (ACOA) is related with the issue of stagnation and Artificial Bee Colony Algorithm (ABCA) is related with the issue of delay in convergence. The same issues are reflected in the usage of these algorithms in finding the optimal solution for Web Service Composition (WSC). A combination of these two algorithms results in the inclusion of salient features of both. Hence an Integrated Ant Colony and Artificial Bee Colony Optimization Algorithm (IACO-ABCOA) based reactive Web Service Composition scheme is propounded considering the merits of transactional and QoS constraints. IACO-ABCOA uses the workflow modeled directed modelled as a directed acyclic graph for solving WSC problem in determining optimal feasible path that aids in estimating the best solution by resolving the stagnation problem of ACOA and results in a better convergence than the rate of convergence enabled by ABCA. IACO-ABCOA is mainly utilized for identifying the near-to-optimal search solution for WSC in an effective and efficient manner in terms of the metrics exploration and exploitation. The experimental results of IACO-ABCOA carried out using a real-world universally available web service inspiring QoS dataset QWS, confirms