Αποδοτικοί αλγόριθμοι και προσαρμοστικές τεχνικές διαχείρισης δικτυακών πληροφοριακών συστημάτων και εφαρμογών παγκόσμιου ιστού

Abstract

In our PhD dissertation we dealt with performance issues in network - centric information systems, netcentric information systems and web information systems. Netcentric approach attempts to depict the augmenting tendency to use the network communication in information systems and web applications in order to provide, to publish, to distribute and to communicate online services and information. The key aim of our doctoral thesis is a) the quality at the service provision, v) the reduction of discovery time and c) the personalization of services and information in network information systems and applications that are based on web engineering technologies. Initially, we studied, designed and implemented efficient algorithms concerning Web Services technologies that have been designed to facilitate interoperable service integration using network infrastructure. Web Services Architecture has been standardized by W3 Consortium (http://www.w3.org) as the technological framework and it has received the wide support of the information technology scientific community as well as the information technology (IT) professionals and industry worldwide. In the first section we introduce a new categorization and comparative presentation of the available algorithmic solutions for service management and discovery. Then, we introduce a series of new efficient algorithms that ensure quality of service provision and improve time complexity in service discovery. Overall in the first part of the thesis we present: - Efficient algorithms for dynamic Web Service selection taking into account non-functional specifications (Quality of Web Service – QoWS) and performance issues during Web Service (WS) consumption attempt (i.e. QoWS enabled WS discovery). - Efficient algorithms for service management and discovery in network centric information systems that are based on decentralized network approaches specifically designed for WS discovery. In the sequel, we propose efficient adaptive methods for personalized web searching. In this way we provide performance improvement both for the internal management and discovery functionality of web based net-centric information systems as well as for the systems’ output that is the end-user information. In particular, in the second section, we introduce a series of three new algorithms for personalized searching. The proposed algorithms are mainly based on link metrics techniques. Their main advantage is that they allow, with the use of a simple methodology, search results personalization, with minimum overhead in terms of storage volume and computation time. We achieve personalized search using link analysis in a personalized graph much smaller one than the whole web graph. The personalized graph is shaped taking advantage of semantic taxonomies. Summarizing the novel research results of this second section are the following: - Efficient algorithms for personalized web information searching. - Adaptive presentation mechanisms of search results with the use of multiple levels of novel categorization. - Extension that allows the adoption of the algorithms for the case of focused web crawling mechanisms, which constitute an alternative personalized searching approach. Finally in the third and last section of our thesis, we present a series of applications, architectures and frameworks of different web based net-centric information environments cases, in which we apply our techniques for service management and personalized information discovery. The main objective of this presentation is to show that the efficient algorithms presented in the previous sections, have multiple potentials of application in problems of different research and technological areas using web based net-centric informative systems and web applications. Cases presented include network management information systems, e-learning approaches, semantic mining and multimedia retrieval systems, web content and structure maintenance solutions and agricultural information systems.