Catálogo de publicaciones - libros

A specification language for performance indicators and their relations and requirements is presented and illustrated for a case study in logistics. The language can be used in different forms, varying from informal, semiformal, graphical to formal. A software environment has been developed that supports the specification process and can be used to automatically check whether performance indicators or relations between them or certain requirements over them are satisfied in a given organisational process.

This paper presents a new Crowded Comparison Operator for NSGA-II to solve the Multiobjective and constrained problem of optimal capacitors placement in distribution systems.

The Partial MAX-SAT Problem (PMSAT) is a variant of the MAX-SAT problem that consists of two CNF formulas defined over the same variable set. Its solution must satisfy all clauses of the first formula and as many clauses in the second formula as possible. This study is concerned with the PMSAT solution in setting a two-phase stochastic local search method that takes advantage of an estimated backbone variables of the problem. First experiments conducted on PMSAT instances derived from SAT instances indicate that this new method offers significant performance benefits over state-of-the-art PMSAT techniques.

In the context of Intelligent Tutoring Systems, there is a potential for adapting either content or its sequence to student as to enhance the learning experience. Recent theories propose the use of team-working environments to improve even further this experience. In this paper an effective matching algorithm is presented in the context of peer reviewing applied to an educational setting. The problem is formulated as an optimization problem to search a solution that satisfies a set of given criteria modeled as “profiles”. These profiles represent regions of the solution space to be either favored or avoided when searching for a solution. The proposed technique was deployed in a first semester computer engineering course and proved to be both effective and well received by the students.

We present an effective method of face and facial feature detection under pose variation in cluttered background. Our approach is flexible to both color and gray facial images and is also feasible for detecting facial features in quasi real-time. Based on the characteristics of neighborhood area of facial features, a new directional template for the facial feature is defined. By applying this template to the input facial image, novel edge-like blob map (EBM) with multiple strength intensity is constructed. And we propose an effective pose estimator using fuzzy logic and a simple PCA method. Combining these methods, robust face localization is achieved for face recognition in mobile robots. Experimental results using various color and gray images prove accuracy and usefulness of the proposed algorithm.

To assess the extent to which individuals adapt themeselfs in a strange cultural environement, the authors analyzed the adaptation process of a number of immigrants who live in Greece. Using categorical variables to represent certain cross-cultural adaptation indicators and employing fuzzy logic clustering, the authors detected and analyzed shifting patterns that are related to the cross-cultural adaptation of individuals.

An algorithmic and formal method is presented for automatic profiling of anonymous internet users. User modelling represents a relevant problem in most internet successful user services, such as news sites or search engines, where only minimal knowledge about the user is given, i.e. information such as user session, user tracing and click-stream analysis is not available. On the other hand the ability of giving a personalised response, i.e. tailored on the user preferences and expectations, represents a key factor for successful online services. The proposed model uses the notion of fuzzy similarities in order to match the user observed knowledge with appropriate target profiles. We characterize fuzzy similarity in the theoretical framework of Lukasiewicz structures which guaranties the formal correctness of the approach. The presented model for user profiling with minimal knowledge has many applications, from generation of banners for online advertising to dynamical response pages for public services.

In the engineering context of control or measurement systems, there is a growing need to incorporate more and more intelligence towards sensing/actuating components. These components achieve some global service related with an intended goal through a set of elementary services intended to achieve atomic goals. There are many possible choices and non-trivial relations between services. As a consequence, both novices and specialists need assistance to prune the search space of possible services and their relations. To provide a sound knowledge representation for functional reasoning, we propose a method eliciting a goal hierarchy in Intelligent Control Systems. To refine the concept of goal with sub-concepts, we investigate a formalization which relies on a multilevel structure. The method is centered both on a mereological approach to express both physical environment and goal concepts, and on Formal Concept Analysis (FCA) to model concept aggregation/decomposition. The interplay between mereology and FCA is discussed.

Ontology merging and alignment is one of the effective methods for ontology sharing and reuse on the Semantic Web. A number of ontology merging and alignment tools have been developed, many of those tools depend mainly on concept (dis)similarity measure derived from linguistic cues. We present in this paper a linguistic information based approach to ontology merging and alignment. Our approach is based on two observations: majority of concept names used in ontology are composed of multiple-word combinations, and ontologies designed independently are, in most cases, organized in very different hierarchical structure even though they describe overlapping domains. These observations led us to a merging and alignment algorithm that utilizes both the local and global meaning of a concept. We devised our proposed algorithm in MoA, an OWL DL ontology merging and alignment tool. We tested MoA on 3 ontology pairs, and human experts followed 93% of the MoA’s suggestions.

Semantic Web relies on Resource Description Framework (RDF). Because of the very simple RDF Model and Syntax, the managing of RDF-based knowledge bases requires to take into account both scalability and storage space consumption. In particular, blank nodes semantics came up recently with very interesting theoretical results that can lead to various techniques that optimize, among others, space requirements in storing RDF descriptions. We present a prototypical evolution of our system called RDFCore that exploits these theoretical results and reduces the storage space for RDF descriptions.