云計(jì)算環(huán)境下個(gè)性化服務(wù)流程生成方法與應(yīng)用研究

摘 要
隨著人們生活水平的不斷提高和個(gè)性化需求的出現(xiàn)，用戶對(duì)服務(wù)內(nèi)容、方式、偏好都在不斷發(fā)展變化，導(dǎo)致傳統(tǒng)的服務(wù)模式已經(jīng)無法很好的滿足用戶需求，人們?cè)絹碓蕉嗟仃P(guān)注于個(gè)性化的服務(wù)體驗(yàn)和服務(wù)品質(zhì)。借助服務(wù)虛擬化技術(shù)利用 Web 服務(wù)組合的方法快速構(gòu)建滿足用戶個(gè)性化需求的服務(wù)流程成為面向服務(wù)軟件工程的一個(gè)挑戰(zhàn)。
然而，在云計(jì)算環(huán)境下存在數(shù)量龐大的 Web 服務(wù)，如何從大量的 Web 服務(wù)中快速、有效、準(zhǔn)確地生成滿足用戶個(gè)性化需求并具有良好用戶體驗(yàn)的 Web 服務(wù)流程成為面向服務(wù)軟件工程的一個(gè)新挑戰(zhàn)。同時(shí)，由于外部 Web 服務(wù)的動(dòng)態(tài)特性，在整合分布式 Web 服務(wù)到現(xiàn)有業(yè)務(wù)流程時(shí)，必須保證當(dāng)服務(wù)發(fā)生異常時(shí)業(yè)務(wù)流程不會(huì)立即終止執(zhí)行。
因此，如何有效地利用現(xiàn)有的 Web 服務(wù)資源，提供一體化的服務(wù)資源訪問機(jī)制，屏蔽服務(wù)多樣性和多變性帶來的系統(tǒng)復(fù)雜度以及動(dòng)態(tài)快捷地構(gòu)建業(yè)務(wù)流程為云計(jì)算環(huán)境下服務(wù)組合提出了新的要求。
針對(duì)上述問題，本文提出了一種面向個(gè)性化需求的分布式 Web 服務(wù)組合框架，用于促進(jìn) Web 服務(wù)的動(dòng)態(tài)整合和異常處理。該框架主要功能包括服務(wù)的搜集、組合、選擇和調(diào)整，通過委托服務(wù)代理為服務(wù)請(qǐng)求者提供一份最佳的 Web 服務(wù)組合方案。該框架能夠監(jiān)控服務(wù)流程的整個(gè)執(zhí)行過程并在單點(diǎn)失效時(shí)根據(jù)用戶反饋及時(shí)進(jìn)行調(diào)整，保證系統(tǒng)整體的服務(wù)質(zhì)量，從而達(dá)到系統(tǒng)的最優(yōu)化。
針對(duì)如何在業(yè)務(wù)層面敏捷構(gòu)建服務(wù)流程、按需創(chuàng)建面向業(yè)務(wù)的云 SaaS 應(yīng)用問題，本文提出了一種服務(wù)流程構(gòu)建方法 TS-Split 算法，該方法利用服務(wù)切片并結(jié)合多目標(biāo)規(guī)劃進(jìn)行求解，同時(shí)給出了服務(wù)切片的推薦算法來解決服務(wù)流程數(shù)量龐大的問題。實(shí)驗(yàn)結(jié)果表明結(jié)合推薦算法的 TS-Split 服務(wù)流程構(gòu)建方法不僅提高了服務(wù)推薦的效率和準(zhǔn)確度，而且降低了獲取需求的難度。
最后，基于上述框架和方法，以旅游服務(wù)為例設(shè)計(jì)實(shí)現(xiàn)了一個(gè)面向個(gè)性化需求的旅游服務(wù)推薦系統(tǒng)。
關(guān)鍵詞：云計(jì)算 服務(wù)組合 個(gè)性化服務(wù)流程 推薦系統(tǒng)
Abstract
With the improvement of people’s living standard continuously and the development of individuation requirements, users of the service content, method and hotspots are in constant development and change, traditional service mode is not very good to meet customer needs. People are increasingly focused on personalized service experience and service quality. Make use of Virtual Service Community Technology virtual reality services as a form of Web services, enabling unified use of the service composition method to quickly build meet the individual requirements of users of the service process.
Therefore, how to quickly, effectively and accurately select and integrate the web services to satisfy the users' personalized need and to supply a good user experience is becoming the new challenge for service-oriented software engineering. Moreover, due to the dynamic characteristics of external web services, while integrating a distributed web service to exiting business process, we must ensure that business process do not immediately come to an end when a service exception occurs.
Thus, how to make effective use of existing Web service resources, to provide integrated services resource access mechanism, to shield service diversity and variability of the system complexity, to dynamic build business process agilely put forward new requirements in the cloud computing environment.
In this paper, we propose an individual demand-oriented distributed framework for web services composition to facilitate the dynamic integration of Web services into business process and fast detection of abnormalities. The major functions of this framework include service collection, composition, selection and adaptation. The framework provides service requester with a best web services composition proposal by deputing services proxy. Meanwhile, the services composition framework monitor the business process execution to adjust immediately when single point fails, to guarantee the service quality of the whole system and optimize it.
One of the key development techniques in the cloud computing environment is to construct agile services, service process at business level, and then create a personalized application according to users' requirements. This paper proposes a new approach named TS-Split to construct the service process. TS-Split uses service slice and multi-objective planning to obtain the optimal service process. Recommendation algorithm of service slice is proposed to solve the problem that the number of service processes is large. This paper focuses on the construction of service process at business level using service slice within time and space constraints. The experimental results show that TS-Split method combining with recommendation algorithm for construction of service process not only improves the efficiency and accuracy, but also reduces the difficulty of requirements acquisition.
Based on the above framework and methods, to tourism services, for example, we designed and implement a travel service recommendation system for user’s personalized demand.
Key words: Cloud Computing Service Composition Personalized Service Process Recommendation System
目 錄
第1章 緒論 1
1.1 研究的背景及意義 1
1.2 國內(nèi)外研究現(xiàn)狀 2
1.3 主要研究內(nèi)容 ..