【摘要】：當(dāng)前服務(wù)計算(SOC)領(lǐng)域的研究重點(diǎn)關(guān)注于Web服務(wù)的發(fā)布、注冊、組織、管理，以及通過服務(wù)發(fā)現(xiàn)、組合等SOC任務(wù)滿足終端用戶的需求。然而隨著互聯(lián)網(wǎng)的發(fā)展和各種計算模式的興起給SOC帶來了新的機(jī)遇和挑戰(zhàn)，一方面，互聯(lián)網(wǎng)上的服務(wù)正在向著大規(guī)模、大數(shù)據(jù)量、多樣性、復(fù)雜性的方向發(fā)展，逐漸形成了以Web服務(wù)為中心的復(fù)雜系統(tǒng)；另一方面，由于服務(wù)之間不能彼此感知，以及缺乏有力的語義支持，使得互聯(lián)網(wǎng)上的服務(wù)存在著信息孤島問題。針對上述挑戰(zhàn)，本文從服務(wù)互聯(lián)分析和社區(qū)發(fā)現(xiàn)的角度對大規(guī)模服務(wù)復(fù)雜系統(tǒng)進(jìn)行相關(guān)研究。 本文是綜合了語義Web技術(shù)、復(fù)雜網(wǎng)絡(luò)技術(shù)、主題建模技術(shù)對服務(wù)計算進(jìn)行的交叉研究。其中語義Web技術(shù)是傳統(tǒng)的，也是當(dāng)前服務(wù)計算研究的重點(diǎn)，它的主要目的是使得服務(wù)具有語義，從而使得計算機(jī)能夠理解服務(wù)、自動化的處理服務(wù)；而傳統(tǒng)的語義Web技術(shù)主要通過借鑒外部知識庫(如本體)實現(xiàn)對服務(wù)的語義標(biāo)注，這就可能導(dǎo)致標(biāo)注不精確以及本體對服務(wù)領(lǐng)域多樣性的適用問題，本文通過借助主題建模技術(shù)學(xué)習(xí)服務(wù)本身蘊(yùn)含的語義信息來進(jìn)一步增強(qiáng)服務(wù)的語義；復(fù)雜網(wǎng)絡(luò)技術(shù)主要針對服務(wù)的復(fù)雜系統(tǒng)問題，從整體上分析面向服務(wù)的復(fù)雜系統(tǒng)的結(jié)構(gòu)、規(guī)律、社區(qū)等。本文主要貢獻(xiàn)如下： 1.面向Web服務(wù)的復(fù)雜網(wǎng)絡(luò)分析和優(yōu)化 服務(wù)形成的網(wǎng)絡(luò)可以有效表達(dá)服務(wù)之間的互操作關(guān)系，而這些互操作是潛在的服務(wù)組合模式�；诰W(wǎng)絡(luò)化的服務(wù)結(jié)構(gòu)，自動服務(wù)組合的過程可以通過圖搜索來完成，從而提高組合的效率�；诜⻊�(wù)之間的互操作關(guān)系，本文自底向上的構(gòu)建了面向語義Web服務(wù)的兩種網(wǎng)絡(luò)模型，并針對網(wǎng)絡(luò)中的孤立節(jié)點(diǎn)和不合理鏈路問題給出了解決方案：通過借鑒網(wǎng)絡(luò)分析，預(yù)測一些對當(dāng)前網(wǎng)絡(luò)具有重要價值的服務(wù)，從而為增強(qiáng)服務(wù)組合網(wǎng)絡(luò)的連通性提供支持；另一方面，針對不合理鏈路，給出一種潛在的基于重用用戶歷史組合案例的組合網(wǎng)絡(luò)優(yōu)化機(jī)制； 2.面向網(wǎng)絡(luò)化Web服務(wù)的社區(qū)結(jié)構(gòu)挖掘 通過研究以Web服務(wù)為中心的社區(qū)結(jié)構(gòu)，能夠揭露網(wǎng)絡(luò)化服務(wù)中深層次的動態(tài)結(jié)構(gòu)，從而為服務(wù)的組織、管理，以及服務(wù)發(fā)現(xiàn)和組合中快速、準(zhǔn)確的定位滿足需求的服務(wù)提供有效的手段。根據(jù)第一部分的網(wǎng)絡(luò)模型，，探討了兩種面向網(wǎng)絡(luò)化服務(wù)的社區(qū)結(jié)構(gòu)：根據(jù)服務(wù)功能參數(shù)在本體中的語義關(guān)系，面向功能語義的社區(qū)結(jié)構(gòu)揭示了具有相似功能的服務(wù)的聚集和競爭現(xiàn)象；基于服務(wù)與服務(wù)之間互操作關(guān)系，構(gòu)建并分析了面向協(xié)作的社區(qū)結(jié)構(gòu)。最后分析了這兩種社區(qū)結(jié)構(gòu)在精化用戶需求和設(shè)計高效組合應(yīng)用上具有較高的價值； 3. Web API/Mashup網(wǎng)絡(luò)分析及挖掘 針對WebAPI及Mashup形成的軟件生態(tài)系統(tǒng)進(jìn)行自頂向下的網(wǎng)絡(luò)建模和復(fù)雜網(wǎng)絡(luò)特性分析。通過構(gòu)建多種網(wǎng)絡(luò)模型對WebAPI及Mashup的功能語義特性以及集成特性進(jìn)行了全面挖掘，分析了所構(gòu)建網(wǎng)絡(luò)的小世界特性、冪律特性以及小集團(tuán)結(jié)構(gòu)，并從實際應(yīng)用的角度探討了這些發(fā)現(xiàn)能夠幫助用戶、服務(wù)提供者有效地利用WebAPI及Mashup生態(tài)系統(tǒng)，如探索式、交互式的用戶需求精化； 4.面向服務(wù)的主題建模及分析 針對傳統(tǒng)服務(wù)語義標(biāo)注不精確性，以及本體對于領(lǐng)域多樣性不適用的問題，探討了主題建模技術(shù)對服務(wù)計算的價值。針對經(jīng)典的WSDL Web服務(wù)進(jìn)行了LDA挖掘，給出了LDA在服務(wù)聚類和社區(qū)劃分中的應(yīng)用；其次探討了本體對監(jiān)督式LDA的借鑒意義，并提出基于本體的Labeled LDA擴(kuò)展模型OEL-LDA，通過實驗對比驗證了所提模型對Labeled LDA中主題質(zhì)量的有效提高，并通過實驗給出了融合本體對精化Labeled LDA中冗余主題的必要性；
[Abstract]:Research in the field of current service computing (SOC) focuses on publishing, registering, organizing, managing Web services, and meeting end-user needs through SOC tasks such as service discovery, combination, and so on. However, as the development of the Internet and the rise of various computing models bring new opportunities and challenges to the SOC, on the one hand, the services on the Internet are developing towards large scale, large data volume, variety and complexity, and gradually form a complex system centered on Web services. On the other hand, because services cannot be perceived to each other, and lack of strong semantic support, there is an island problem of information on the Internet. Aiming at the above challenge, this paper studies the complex system of large-scale service from the view of service interconnection analysis and community discovery. This paper deals with the intersection of semantic Web technology, complex network technology and topic modeling technology on service calculation. Research. Semantic Web technology is traditional and is the focus of the current service computing research. Its main purpose is to make the service have semantic, so that the computer can understand the service and automation. The traditional semantic Web technology can be used to mark the semantic annotation of the service by referring to the external knowledge base (such as ontology), which can lead to inaccurate marking and the application of ontology to the service field. In this paper, the semantics of service is further enhanced by learning the semantic information contained in the service itself with the help of the topic modeling technology, the complex network technology mainly focuses on the complex system problem of the service, the structure, the law and the society of the complex system facing the service are analyzed from the whole. Area, etc. The main contribution of this article The following: 1. Complex Network for Web Services The networks formed by the network analysis and optimization service can effectively express the interoperation relationship between services, and these interoperation are latent in a networked-based service architecture, the process of the automated service combination may be accomplished through the graph search, from In order to improve the efficiency of the combination, this paper constructs two kinds of network models facing semantic Web service from the bottom up, and gives the solution to isolate nodes and unreasonable links in the network. By using the network analysis, we can predict some services which have important value to the current network, so as to provide support for the connectivity of the enhanced service combination network; on the other hand, aiming at the unreasonable link, a potential reuse user history combination case is given. Combinatorial network optimization mechanism; 2. network oriented Web service community structure mining can expose the deep dynamic structure in the network service by studying the community structure centered on Web service, thus serving the organization and the pipe of the service. Quick and accurate positioning in discovery and combination of services According to the network model of the first part, two community structures for networked services are discussed: the semantic relation of the service function parameters in the ontology, and the community structure facing the functional semantics reveals the phase of the community structure. The aggregation and competition of functional services; interoperation between services and services. The collaborative community structure was built and analyzed. Finally, the requirements and design of the two community structures in refined users were analyzed. High-efficiency combined applications with higher value; 3. 1.3. 1. b API/ Mashup Network Analysis and Mining Software Ecosystem for WebAPI and Mashup In this paper, a top-down network modeling and complex network characteristic analysis are carried out. By constructing a variety of network models, the function semantic characteristics and integration characteristics of WebAPI and Mashup are comprehensively mined, and the small world characteristics, power law characteristics and small group structure of the constructed network are analyzed. The perspective of practical applications explores these findings to help users and service providers effectively leverage the WebAPI and Mashup Ecosystem, such as exploration, interactive User Requirements refinement; 4. Service-oriented topic modeling and analysis of traditional services The semantic annotation is not accurate, and the ontology is much in the field. This paper discusses the value of the subject modeling technology to the service calculation, discusses the application of the ontology to the service and community partition, and then discusses the significance of the ontology to the supervised classification, and puts forward some suggestions. Based on the body-based Labeled Dirichlet extended model OEL-044, the effectiveness of the proposed model on the quality of the subject in Labeled has been verified by experimental comparison, and the experimental results are given.
【學(xué)位授予單位】：天津大學(xué)
【學(xué)位級別】：博士
【學(xué)位授予年份】：2014
【分類號】：TP393.09

【參考文獻(xiàn)】

相關(guān)期刊論文 前2條

1 鄧水光;吳朝暉;;Web服務(wù)組合方法綜述[J];中國科技論文在線;2008年02期

2 楊春芳;;圓頂閥在固體物料傳輸中的應(yīng)用[J];中國儀器儀表;2010年10期

本文編號：2311994