Research On Key Technologies Of Internet Of Service Ecosystem Evolution And Assistant Decision Supporting

With the rapid development of new technologies such as cloud computing,edge computing and mobile computing,the number and diversity of available services has exploded.The rapid growth of cooperation and competition among services has given rise to the new phenomenon of "Internet of Services".The "Internet of Services" has given birth to a large number of new forms of services that can provide users with unique value that the original individual services could not provide.The Internet of Services is an open ecosystem that requires in-depth knowledge and scientific governance to enable the healthy and sustainable development of all participants in the Internet of Services.Thus,it is particularly important to understand the architecture,evolution mechanism,and scientific use of the Internet of Services.However,existing works lack in-depth and systematic research on the Internet of Services ecosystem.Firstly,the existing works have insufficient depth in characterizing the evolution of the Internet of Services ecosystem and lack corresponding model construction methods,which cannot clearly and effectively elucidate the basic components of the service ecosystem and the driving mechanisms of its evolution.Secondly,the existing works have a single observation perspective and insufficient intelligence in their analysis methods when analyzing the evolution of the Internet of Services,resulting in inaccurate evolution analysis results that are difficult to further utilize.Finally,the Internet of Services is an open ecosystem with technological-business duality that continues to evolve.Existing works tend to focus on software technology services when using the Internet of Services to assist decision-making and do not fully consider the evolutionary characteristics of the Internet of Service,leading to underestimation of its application potential.In response to these challenges,this paper takes an ecological perspective and conducts research on the evolution model of the Internet of Services ecosystem,the evolutionary analysis of the Internet of Services,and the decision support for the Internet of Services.(1)To address the issues of rough models and missing modeling methods in existing research on the evolution of the Internet of Service ecosystem,this paper proposes a service ecosystem model based on a multi-layer dynamic semantic network.The hierarchical structure of the ecosystem model provides a comprehensive and accurate description of various elements in the service ecosystem,introducing the concept of service events to provide a good explanatory mechanism for the evolution of the service ecosystem.A data-driven service ecosystem modeling method is proposed,which realizes the modeling process based on active learning and natural language information extraction,achieving the goal of automatic/semi-automatic large-scale service ecosystem modeling from massive open data at low labor costs.(2)To address the issue of a single level in the analysis of the evolution of software service ecosystems,this paper proposes a multi-level service ecosystem evolution analysis method based on the stochastic block model.A mathematical model that can support indepth evolution analysis at three levels,including the overall service ecosystem,service communities,and individual services,is presented.The basic evolutionary laws of the software service ecosystem are summarized.(3)To address the issue of insufficient intelligence in analyzing the evolution of business service ecosystems,this paper proposes a service ecosystem evolution analysis method based on tracking the evolution of service communities.This method takes a unique perspective of service communities to analyze the evolution of the ecosystem and tracks the evolutionary trends of service communities in the Service Internet ecosystem,explaining the driving factors that lead service communities to evolve in different directions.(4)To address the problems of narrow application scope and inadequate consideration of service ecosystem evolution characteristics in the existing research of service ecosystem assistant decision-making.This paper studies decision support methods for typical Service Internet scenarios targeting different roles.This method greatly improves the performance of existing models in decision support scenarios and expands the scope of decision support for Service Internet.Specifically,this paper considers service mashup creation scenarios,potential cooperation partner recommendation scenarios,and service event prediction scenarios.A service aggregation recommendation method based on dynamic representation learning and alignment is proposed,providing a set of service solution recommendations for specific needs through an end-to-end model implementation based on dynamic graph neural networks,reducing the burden on software developers.A potential cooperation partner recommendation method based on fine-grained behavior perception is proposed,which introduces fine-grained behavior characteristics in the evolutionary process of the service ecosystem as additional inputs to help select potential cooperation partners using dynamic graph neural networks.A service event prediction method based on dynamic graph neural networks is proposed,which not only improves the accuracy of predicting future events but also accurately estimates the time of event occurrence,ultimately helping regulators formulate strategies in advance to cope with various events.In conclusion,this paper focuses on the Internet of Services from an ecosystem perspective,conducts an in-depth discussion on the three major research questions of ecosystem modeling,evolution analysis and assistant decision-making,proposes corresponding theories and methods,and lays the foundation for a comprehensive understanding and scientific utilization of Internet of Services ecosystem.