Catálogo de publicaciones - libros

In this paper we discuss how knowledge integration throughout system analysis, modelling and development courses can be stimulated by giving an overview of our MIRO-project at K.U.Leuven. This includes offering an online knowledge base of all-embracing case studies, structured according to the Zachman framework. Supported by collaborative groupware, students not only get the opportunity to consult and compare solutions for the case studies, but also actively discuss and contribute to alternative solutions. In this Problem Based Learning (PBL)-context, students are able to influence and understand the development of a certain process through interactive computerized animations and demos.

In this paper we present some ideas, concepts and experiences related with our work of implementing solvers to combinatorial optimization problems. We have been designing, implementing and using solvers for important kind of problems, and painfully, the process of developing them is not optimal at all. The most used practice, which is based on trial and error, often incurs an enormous amount of resources. Clearly, this is a Software Engineering problem, then to find better ways of developing optimization algorithms, solvers and metaheuristics is our interest too. The software development community has a wide spectrum of methodologies when it decides to implement a software project. From the more traditional Tayloristic practices to Agile methods. Software development is a knowledge intensive activity and the knowledge creation and sharing are crucial parts of the software development processes. This paper presents a comparative analysis between knowledge sharing approaches of Agile and Tayloristic software development teams sprinkled with concerns about the development of Metaheuristics.

In this paper we propose a framework for predicting access patterns for materialized methods. To this end, we analyze past access patterns and compute frequencies of method calls. Based on these frequencies, the system automatically decides when to recompute materialized methods. The framework was implemented and experimentally evaluated. The results are reported in this paper.

In this paper a software architecture design process based on a supportable meta-architecture (SMA) and roundtrip engineering is proposed for large software projects. Our process is applied after the requirements elicitation and analysis phases of a software project. The process begins with designing a SMA aimed at minimizing and managing software complexity. The meta-architecture should be highly supportable, i.e. understandable, maintainable, scalable, and portable and based on software engineering principles particularly object oriented design techniques, design patterns and frameworks. Roundtrip engineering embraces various supportability metrics to ensure that the implementation conforms to the meta-architecture and that the resulting system at the end of each development iteration period is supportable. Two project case studies using this design process are also presented in the paper.

The Knowledge-Based Enterprise Modelling framework is based on the redefined concept of the knowledge-based enterprise. The Strategic Alignment Framework by Henderson and Venkatraman is modified and used for description of Knowledge-based Enterprise structure. The Enterprise Knowledge Base together with explicitly defined knowledge management activity is treated as the major component of the knowledge-based enterprise. Four domains of the knowledge-based enterprise are abstracted from the enterprise architecture point of view. The process-oriented Knowledge-based Enterprise Model is presented by modifying the Porter’s Value Chain Model (VCM). The Knowledge management layer is identified within the modified VCM. Two types of control loops are identified between layers of the Knowledge-Based Enterprise Model.

Recently many researchers have focused on management of historical trajectories of moving objects due to numerous size of accumulated data over time. However, most of them are concentrated in Euclidean spaces with (, , ). In real world, moving objects like vehicles on transportation networks have constraints on their movements, and some of applications need to manage and query them. Previous work based on Euclidean is inefficient to process trajectories on road networks. In this paper, we propose a indexing method for trajectories of moving objects on road networks. While some work has been done for indexing the trajectory in spatial networks, little indexing method support the network-based spatiotemporal range query processing. Our method consists of multiple R-trees and graph structures to process the network-based spatiotemporal range query defined by the network distance instead of Euclidean distance. Consequently, we show that our method takes about 30% less in node accesses for the network-based range query processing than other methods based on the Euclidean distance by experiments.

In Constraint Programming, enumeration strategies are crucial for resolution performances. The effect of strategies is generally unpredictable. In a previous work, we proposed to dynamically change strategies showing bad performances, and to use metabacktrack to restore better states when bad decisions were made. In this paper, we design and evaluate strategies to improve resolution performances of a set of problems. Experimental results show the effectiveness of our approach.

Security of information systems is an increasingly critical issue. Access control is a crucial technique ensuring security. It should be based on an effective model. Even if some approaches have already been proposed, a comprehensive model, flexible enough to cope with real organizations, is still missing. This paper proposes a new access control model, FORBAC, which deals with the following issues: The first one is the adaptability to various kinds of organization. The second one concerns increasing flexibility and reducing errors and management cost, this is done by introducing a set of components which allow fine-grained and multi-level permission assignment. The paper introduces a framework for evaluating the proposed approach with respect to other related research through views, facets and criteria.

A novel anycasting protocol for timed asynchronous distributed multiserver systems is proposed. It is based on an anonymous multicast communication in a network of servers. In contrast with known approaches, the task of selection of a server in the given group is shifted from clients and the network to the servers that make a selection decision themselves as a result of a negotiation. The protocol was investigated with the use of a detailed simulation model using a class of the extended Petri nets. The results of simulation study of the proposed protocol are compared to the behavior of an ideal, centralized multiserver queuing system.

This paper presents a scheduling scheme for wireless broadcast data which considers data access frequencies. Whereas the previous approaches consider either data replication or index replication, the proposed scheme considers both of them. Through analysis and experiments, it is shown that the new approach has better performance (with respect to the access time and tuning time) than the previous approaches.