Font Size: a A A

Research On Method And Key Technologies Of Task-oriented Military Information Service Composition

Posted on:2012-06-11Degree:DoctorType:Dissertation
Country:ChinaCandidate:X H YuFull Text:PDF
GTID:1118330341951625Subject:Military Operations Research
Abstract/Summary:PDF Full Text Request
SOA/Web Service technologies have the characteristics of loosely coupled,supporting efficient integration of application systems and accommodating changes ofbusiness processes. So they have become important supporting technologies to realizethe sharing of military information resources and integration of applications in gridenvironment. As the development and deployment of more and more MilitaryInformation Services(MIS) and considering both the decomposable, sectional attributesof operational missions and the limitation of MIS's functional granularity,Task-Oriented Military Information Service Composition(TOMISC) will become acommon phenomenon in constructing and applying Service Based Military InformationSystems(SBMIS). Unfortunately, due to the differences of applying area, there are stillmany challenges when directly using the business field's ordinary Web servicecompositiontechnologies.Withthesupportofnationaladministration'sresearchproject,this thesis studies four aspects of TOMISC: conception and framework, servicedescription, finding and selection, service composition modeling, and servicecomposition optimizing. The aim of this study is to break through key technologies toconstruct flexible and effective military information system and to realize the requiredsharingandintegrationofmilitaryinformationresources.Maincontentsandaccomplishmentsofthisthesisareasfollows:(1)The concept and framework related to TOMISC are built. Firstly, thesuperiorities of SBMIS and the differences between SBMIS and enterprise SOA basedsystems are analyzed. Secondly, on the basis of analyzing the relationships of task, C2organization,C2resource, battleaction andMIS,thedefinitionofTOMISC is proposed.Then, the TOMISC's working pattern, system integration framework, protocols andMIS categories are studied. Finally, the process of service composition in TOMISC isproposed to further determine the relationship of key technologies needed to be solvedinthethesis.(2)A non-functional attribute based MIS description, discovery and selectionmethod is studied. Towards the domain characteristics of MIS, the behavior restrictionmodel that describes the access, context of MIS and the QoS model that describes theMOC,MOPandMOEofMISareproposed.Afterthat,aservicefilteringmethodbasedon behavior restrictions in service discovery and a service selection method based onsimulation of Object Petri Net are proposed. In service discovery, referring to the levelbased matching idea, a similarity matching arithmetic of behavior restrictions ispresented for service filtering and discovery. Case analysis shows that the method canefficiently improve the precision with a certain guarantee recall by setting reasonablefiltering threshold. In service selection, the thesis studies the corresponding relationships betweenservicecompositionmodelandObject Petri Net inatomicservice,three-dimensional QoS and service composition logic. Then, the application of ObjectPetri Net in guiding the establishment of service selection metrics and in analyzingservice'sQoSisanalyzedtoproofthefeasibilityandvalidityofthemethod.(3)A multi-view based executable modeling method of MIS composition isproposed. Towards the limitation of single modeling technology and the requirementsof modeling in MIS composition, and also referring to the model driven Architecture'smodel mapping idea and the executable Architecture's supporting model simulationviewpoint, a multi-view based executable framework is proposed. Then the modelingand related model conversion method in different views of the framework is studied.The proposed method achieves the phase description and smooth transformation ofdifferent models from task requirement to service composition modeling, simulationand realization. This can both ensure the agility of TOMISC and significantly improvethe accuracy and feasibility of service composition modeling. The thesis also analyzesthe supports of the proposed related model towards DODAF's Service View productsandtheexecutablearchitecture.(4)A Parallel distribution based optimization method of service composition isstudied. Towards the timeliness and successful probability requirements of operationaltask in MIS composition and by analyzing the optimization problem of MIScomposition,aparalleldistributionframeworkthatsupportservices'dynamicbindingisproposed. Based on analyzing the relationships of the allotted services, the completedtime and the successful probability of a single service node, a mathematicalprogrammingservicedistributionmodel forcompositedservices is established.Thenanimproved particle swarm optimization algorithm that introduced random disturbance,searchingboundaryoptimizationandsurvivalofthefittestswarmisproposedtoresolvethe mathematical model. Experimental results show that parallel distribution method ofservice resources and the optimal algorithm are effective methods to efficientlydistribute service resources and ensure the timeliness and successful probabilityof MIScomposition.In the end, based on the research of above theories and keytechnologies, the mainfunctional modules, prototype systems and key services that related to TOMISC aredesigned and realized. The case application in missile defense background validates thefeasibility, effectiveness and correctness of the proposed theories, methods and theconstructedsystemsinthethesis.
Keywords/Search Tags:MilitaryInformationService, ServiceBasedMilitaryInformationSystem, Service Composition, Service Discovery and Selection, ServiceComposition Modeling, Service Composition Optimization
PDF Full Text Request
Related items