A Heuristic Algorithm For QoS (Non-Functional) Based Service Matching

The problem of QoS-based selecting a web service dynamically or composing a set of web services to conduct a business task has been investigated in this paper. It outlines the discovery of either atomic or composite services satisfying the request QoS requirements. After, the functional requirements matching is achieved where the request Inputs/Outputs syntactically and semantically match services signature, there may exist more than one matched service, then the QoS (nonfunctional) attributes best fit service is chosen. The QoS may be set by the service consumer or computed from past executions. If the chore cannot be satisfied by atomic process, it should be bind with a virtually composed service (set of atomic services).A collection of services sets may be selected where each set has a number of atomic services assembled together to result the desired business requirements. The best fit set of services based on QoS parameters is selected. This paper only concentrates on QoS parameters based selection. The problem is formulated as an optimization problem, so an optimization solving techniques such as integer programming can be applied. A heuristic technique is designed and evaluated.