| At present,a variety of Web services are emerging on the Internet,and the business process of the service system is becoming more and more complex.The number of tasks and the number of candidate services in the business process are increasing exponentially,resulting in the rapid expansion of the scale of the service composition scheme.Facing the increasingly complex service needs of users and enterprises,how to adapt to the new challenges of large-scale complex service environment and how to quickly return high quality composite services to meet the needs of users and enterprises within a reasonable time limit have become the key problems to be solved in the field of service computing applications.Based on the analysis and summary of the existing large-scale Web service composition methods,this thesis proposes a large-scale Web service composition framework,and deeply investigates the problem of large-scale Web service composition from two aspects: large-scale Web service filtering and large-scale Web service composition modeling and solving.In the aspect of large-scale Web service filtering,in order to reduce the service composition space and improve the accuracy of service filtering,a Web service filtering method based on DSSM model is proposed.Firstly,the Skip-thought model is used to obtain the vector representation of service function requirement description and Web service documents;Then,the semantic features of the vector of service function requirement description and Web service document are extracted based on DNN model;Next,service filtering is carried out according to the semantic similarity between the two,and the Top-k candidate services that meet the functional requirements are filtered for each abstract service to form the candidate service set of service composition.In the aspect of modeling and solving large-scale Web service composition,in order to obtain high quality service composition scheme and shorten service composition time,a Web service composition method based on QoS and improved Gray Wolf algorithm is proposed.First,the QoS-based large-scale Web service composition model is constructed;Then,by introducing chaotic mapping and nonlinear convergence factor,the Gray Wolf algorithm is improved to enhance the global and local search ability of the algorithm and reduce the risk of falling into local optimization;Finally,under the Spark framework,the improved Gray Wolf algorithm is used to solve the QoS-based large-scale Web service composition model to generate the optimal service composition scheme.The experimental results show that this method has obvious advantages in the quality and time efficiency of large-scale Web service composition. |
