An Improved Genetic Algorithm-Based Web Service Composition
Hao Long · Advanced materials research · 2011
In this paper a novel web service composition approach based on improved genetic algorithm is proposed. Firstly the relative vector is defined to represented the input and output parameters of user requirements and web services, by comparing the relative vectors, a composed service can be evaluated how much it match user requirements, thus the web service composition is transferred as a linear programming problem. Secondly an improved genetic algorithm is used to search optimum composite services. Compared with the existing approaches, this method supports dynamic service composition, bears component services with multiple input & output parameters, and enables inclusion of parallel and choice structure inside the composite services.