Gábor Magyar

Cost-optimal topology planning of hierarchical access networks

This paper deals with the problem of cost-optimal hierarchical topology planning for telecommunications access networks. The generalized problem is computationally hard and only its substantially simplified variants have been studied before. We propose a heuristic algorithm relying on iterative problem decomposition, clustering methods and local optimization. The main idea of the approach is to combine clustering and local optimization operators and to apply the combined operators iteratively within the levels of the hierarchy. By an extensive empirical analysis it is shown that the presented method efficiently solves the underlying optimization problem for instances of practical size, furthermore, it is robust to handle flexibility in the problem statement and in the cost function.

Reroute sequence planning for label switched paths in multiprotocol label switching networks

This paper deals with the problem of reroute sequence planning for label switched paths (LSPs) in multiprotocol label switching (MPLS) networks. This problem arises when the path set is recalculated by a global optimization method for achieving a better resource utilization. In this case we have to find a sequence of the one-by-one reconfiguration of the LSPs from their original paths to the optimized paths, with the constraint that the link capacities should not be violated at any time during the reroute process. The problem is computationally hard in real-world situations because of the large amount of routers and LSPs in the network, so that neither exact nor approximation algorithms have practical relevance. We introduce a number of heuristic approaches, which are investigated through extensive empirical tests based on real network topologies. The results reveal the complexity of the problem. We propose possible solutions and suggest future research directions for alternative answers to the reroute sequence planning.

Design of survivable VPN based VoIP networks

This paper addresses the issue of survivable voice over IP (VoIP) network design. In the applied model the whole VoIP network is divided into two logical components: the access network and the transport network. The access network consists of VoIP end-points, which connect to the transport network through edge routers serving as gateways. Thus, one task of the design process is to assign gateways to the VoIP end-points. To assure security and availability for the VoIP traffic, a virtual private network (VPN) is assumed as transport network in this paper; therefore, the other task is to specify the transport VPN in the most economical way, while also taking the possible failures of transport network elements into consideration. Assuming a single failure at a time, the survivability of the VoIP network can be assured by assigning two gateways, namely a primary and a backup one, to each VoIP end-point, and calculating two disjoint paths between each pair of gateways. The two tasks of survivable VoIP network design can be solved separately; however, in this paper, a novel solution is proposed in which the edge router assignment process takes both the objective function of VPN specification and the issues of survivability into consideration. Multiple methods realizing the novel approach are presented, which are based on the paradigms of evolutionary algorithms and simulated annealing. The performance of the presented methods is evaluated with the help of simulations, using a well-known greedy algorithm as reference. It is shown that the proposed methods outperform the reference algorithm significantly in the simulation scenarios investigated.

Entity recognizer in hungarian question processing

In our ongoing research and development project, called “In the Web of Words” (WoW), funded by the National R+D Program in Hungary, we aim to create a complex search interface that incorporates—beside the usual keyword-based search functionality—(1) deep web search, (2) Hungarian natural language question processing, (3) image search support by visual thesaurus. This paper focuses on a particular and crucial part of the question processing problem (2): recognition of entities. Entities are expressions that have fixed form, and that are assigned context specific information in the dictionary. Due to the agglutinative feature of Hungarian language they often appear in the text differently as in the dictionary, therefore their detection requires special algorithms at processing.

A Fuzzy Information Propagation Algorithm for Social Network Based Recommender Systems

Web-based services that have become prevalent in people’s everyday life generate huge amounts of data, which makes it hard for the users to search and discover interesting information. Therefore, tools for selecting and delivering personalized contents for users are crucial components of modern web applications. Social recommender systems suggest items to users assuming the knowledge of the users’ social network. This new approach can alleviate the common weaknesses of traditional recommender systems, which completely ignore the users’ personal relationships in the recommendation process. In this paper, a social network based fuzzy recommendation technique is presented, which propagates information through the users’ social network and predicts how users would probably like a certain product in the future. Experimental results on a public dataset show that the proposed method can significantly outperform popular and widely used recommendation system methods in terms of recommendation coverage while maintaining prediction accuracy and performs especially well for cold start users, that have only rated a few items or no item at all previously.

Building an Open Document Management System with Components for Trust

Nowadays the Web can be considered as the largest database or more precisely, the largest document archive ever built. None the less the accessible information resources are in the deepness (principally in structured databases), and cannot be accessed by a simple free text search, the agents of Google index more than three thousand million pages. As technology and networking becomes more and more accessible and affordable, the number of documents increases exponentially.

The Impact of Technological Paradigm Shift on Information System Design

The evolution of ISD methods from water flow model to object-oriented techniques was catalyzed by the need of managing several uncertainties in user requirements, changes in software and hardware technology as we have reported in our previous case study 1. However it was assumed, that there are only slight changes in the static and dynamic behavior of the completed system during operation.

Information System Development in the “E-Age”

Audio-visual content plays increasing role in cultural and scientific life, however preservation and accessibility methods of such objects is still under research and development in all over the world. The problem arises from the multidisciplinary and heterogeneous nature of content on the one hand, and the continuos and high speed change in infocommunication technology on the other hand. Successful co-operation of several different kinds of multicultural, multilingual repositories require the use of widely accepted open standards and systems, however tailored collaboration among the project participants, the institutes and experts of information technology, media and regulation is an exciting challenge for project management.

Which players will leave their community? Predicting guild abandonments in world of warcraft game data

World of Warcraft (WoW) is one of the most popular massively multiplayer online role-playing games (MMORPGs) having more than 10 million subscribers over the world. In order to engage and retain users understanding and predicting their behavior can be very useful for game developers. An important component of WoW are so-called guilds, which are social communities whose members can act together efficiently to accomplish more difficult goals and also provide a social atmosphere in which the game might be more entertaining. In this paper, we build predictive models to forecast which of the players will leave their guild in the close future. Our best model uses fuzzy c-means clustering to capture groups of similar guilds, that serve as the basis of an ensemble model, which computes predictions for each cluster separately and combines individual predictions into one final prediction using the memberships of the fuzzy clusters. Empirical analysis on WoW game data shows that our methods convincingly outperform the only existing method in the literature. To ensure transparency and reproducibility we publish the source codes of the research and also provide a Docker image, which makes it possible for anyone who has Docker installed to reproduce all of our results with a single command.

Collaborative Media Content Management

The paper presents the design and implementation of an original integrated media content management system. In this very fast developing area where numerous changes occur even in a year’s time it is really hard to foresee future trends and successful directions. The primary goal was to develop an example of a Collaborative Media Content Management System where different media types and other online contents can simply be stored integrated in one system and can easily be managed and shared by users and communities as well. Right at the beginning, we were aware of the fact that the project would last one year. Therefore, it was not enough to include the then newest features in the plan: it was also necessary to predict future trends and requirements, so that by the time the project is finished, it would still be innovative. The functional design of the application was based on the observation of new internet user generations who are heavily involved in real-time messaging and community portals. Based on our innovative ideas the system has been finished and besides our successful project many similar solutions have been appeared all over the world. The paper presents the most important features of our approach and the outcome: a Collaborative Media Content Management system, called simply Media-store, and emphasizes the innovative services of the system (visual tagging, mediashow, mediablog, common media arrangement).