Research On Web Services Composition Method Based On Semantic Similarity

More and more web services have been published in the Internet in recent years and there are generally huge amounts of services available when users try to build a coarse–grained service by utilizing some fine–grained ones.Therefore,how to composite multiple simple services to meet new requirements is becoming a heated issue.On the basis of the existing web service composition technologies,this thesis proposes a web service composition method based on semantic similarity of web services.First of all,the two ways of web service description were studied,namely the semantic service description and the syntactic service description.Because the semantic way supports the intelligent planning better compared to the traditional syntactic way.This thesis took the web service ontology Language OWL-S as the web Service description tool.Secondly,the traditional algorithms,such as the semantic similarity algorithms based on distance and based on the amount of information were analyzed,including the comprehensive algorithm that took the two factors into consideration at the same time.In the process of service matching,the different matching direction can result in different precision.On the basis of the above mentioned comprehensive algorithms,this thesis puts forward an improved method of calculating the semantic similarities,which introduces the dissymmetry factor.Then by adopting the proposed semantic similarity algorithm,the similarities of web Services are calculated,which are used in the clustering algorithm,generating clusters and each cluster respectively represents a function.Thus,the matching accuracy is improved and the effects of service composition are also optimized.Finally,combining with the hierarchical task network planning engine,a dynamic service composition model is proposed.Its implementation contains several core modules,such as service clustering,planning and compositing,service matching and execution.The results show that the web service composition method based on semantic similarity has certain superiority and application value,compared to the traditional one.