Research On Web Service Composition Based On Improved Crow Search Algorithm

With the increasing demand of People’s Daily life,the Internet platform has produced many emerging Web applications.A complete Web application can consist of Web services distributed in different clouds.These Web services distributed in different clouds have many candidate services with similar functionality but different service quality to choose from.How to combine these candidate services accurately and efficiently becomes the key to effectively save resources and improve user satisfaction.Traditional optimization methods usually have complex coding,many parameters and low optimization efficiency when solving the problem of service composition.Therefore,this thesis introduces an intelligent optimization algorithm – crow search algorithm,which has the advantages of fewer parameters,easy to understand and fast convergence speed.Although the crow search algorithm has good optimization performance,it has some disadvantages such as fixed parameters,low solving precision and poor global searching ability.So in view of the problems existing in the crow search algorithm,this thesis first puts forward an adaptive strategies of perceived probability,the perceived probability can be obtained at the early stages of the iterative relatively large values,ensure early has good global search ability,along with the iteration,perceived probability decreases,late guarantee good local optimization ability.Secondly,some dimensional components of the individuals in the crow search algorithm may be trapped in the local optimum of the current dimension.Therefore,this thesis proposes an optimal individual updating strategy,which variants the optimal individuals according to the relative change rate of the dimensional components of the optimal individuals of two adjacent generations.In this thesis,we first test the performance of the improved crow search algorithm on 12 classical benchmark functions.The results show that the optimization accuracy and convergence speed of the proposed algorithm are better than those of other algorithms.Then,this thesis designs a Web service composition prototype system based on the improved crow search algorithm,which simultaneously integrates several classical algorithms of the same type.Finally,an actual case shows that the improved crow search algorithm in the prototype system has better performance in solving the problem of Web service composition.