Certain investigations on qos driven web service composition using metaheuristic technique

Abstract

Service Oriented Computing (SOC) has been widely applied in academic institutions and industries to design adaptive software applications in a loosely coupled manner. One of the main aspects of service composition is to develop higher-level services that are called composite services. The common issues that arise in SOC are automatic assembling of web services and data heterogeneity among web services that take part in the composition. In distributed environments, the services are used without any quality guarantee from the service provider, and so it can negatively affect a composite service. The present study proposes a solution to the problem of composition of web services on the basis of non-functional parameters including customer focused Quality of Service (QoS) parameters such as response time, cost, reliability and availability. During composition, the service considered would satisfy the customer s expectations. Three major contributions such as Hybrid Particle Swarm Optimization (HPSO), Fuzzy Hybrid PSO (FHPSO) and Map Reduce Fuzzy Hybrid PSO (MR-FHPSO) are proposed to obtain efficient workflow with high QoS in the web service composition with minimum execution time. The first phase of the research involves the employment of the metaheuristic methods based on PSO and Genetic Algorithm (GA) to solve the optimization issues. QoS based web service composition is done by utilizing HPSO approach. Three different hybridization approaches of PSO with GA such as Hybrid PSO Type I, Hybrid PSO Type II and Hybrid PSO Type III are proposed. The search space of web service composition consists of a largenumber of composed services. newline