Land Cover Dynamic Service Composition Methods Based On Complex Network Theory

Land cover and its changes information are essential basic data and key parameters in studies related to natural resource monitoring,climate change analysis and global sustainable development.With the development of the Internet and big data technology,organizations have been disseminating and publishing land cover data and algorithms on the Internet.However,users cannot meet their complex land cover task needs with just one land cover service.To meet users’ higher level needs,multiple types of single land cover service must be orderly combined into a land cover service chain.As a consequence,a key scientific problem in the field of land cover is how to make full use of existing open and massive land cover service resources and build a dynamic,efficient,and accurate land cover service chain on-demand.Various Web service composition methods have been proposed at home and abroad,such as artificial intelligence planning,genetic algorithms,ontology semantic reasoning,and loose programming.These methods provide ideas for the land cover service dynamic composition.However,existing methods struggle to cope with complex and variable land cover tasks.The reason for this is that existing studies still lack in-depth exploration of land cover service relationships,heterogeneous land cover service invocation,and service adaptive composition.To address this problem,this thesis proposes a dynamic composition method of land cover service based on complex network.The basic idea is as follows: first,the land cover service weighted complex network is constructed by the semantic invocation relationship among the land cover services.Then,based on Graph Plan,composite services are dynamically searched from the service complex network.Finally,the land cover service composition adaptive mechanism is used to repair the abnormalities that occur during the land cover service dynamic composition.This is done to ensure the smooth execution of the land cover service composition.The main research is as follow:(1)Land cover service weighted complex network modelingTo address the problem that land cover services in cyberspace are isolated and lack association relationships,this study proposed a weighted complex network model for land cover services and constructed the complex network via semantic invocation relationships between land cover services.The calculation of the edge weights between the nodes of the land cover service was defined to indicate the closeness of the connection between the services.The study also analyzed the model update method to accommodate the dynamic evolution and development of the service network.(2)Knowledge-driven land cover service dynamic composition methodTo address the problem of difficult utilization of heterogeneous services and strong a priori knowledge dependence of the land cover service composition,the study proposed a knowledgedriven land cover services dynamic composition method.First,a land cover service ontology consisting of a task ontology,a service ontology,a service chain ontology and a data ontology is proposed.This is to provide a semantic description of the objects required for land cover service composition through a unified standard.Then,the service composition problem was also considered a graph planning problem.A constraint-aware algorithm was proposed,which incorporates prior knowledge expressed as knowledge constraints into the algorithm.Its objective is to improve land cover service composition accuracy by dynamically searching for services in a complex network of land cover services.(3)Land cover service composition adaptive mechanismAiming at the problem of insufficient abnormal adjustment capability of the land cover service composition method,the study combines the characteristics of the service complex network and the land cover domain,analyzes the adaptive framework,the anomaly classification system,and the anomaly decision method,and proposes an event-driven adaptive model based on the land cover service composition.The model realizes the land cover service composition adaptive adjustment process of anomaly sensing,anomaly analysis,anomaly decision,decision execution and service network update based on the existing knowledge,which guarantees the practical availability of land cover service composition technology.(4)Development and implementation of land cover service composition prototype systemBased on the above theoretical research,the prototype system of land cover service composition(ILC-LCSC)was developed with the help of Web GIS,service computing and other technologies,which integrated the weighted land cover service complex network model,the constraint-aware land cover service dynamic composition and the land cover service composition adaptive mechanism into the system module.An example of land cover change detection service composition was developed.The user only needs to provide input data and desired output,and the system can dynamically build a service chain for land cover on demand.It also supports the detection and repair of service composition anomalies.The experiment proves that this study provides a complete solution to the problem of the dynamic composition of land cover services.It realizes the on-demand processing of land cover information by Web service.