Fernando Sáenz

A hybrid evolutionary approach for solving constrained optimization problems over finite domains

A novel approach for the integration of evolution programs and constraint-solving techniques over finite domains is presented. This integration provides a problem-independent optimization strategy for large-scale constrained optimization problems over finite domains. In this approach, genetic operators are based on an arc-consistency algorithm, and chromosomes are arc-consistent portions of the search space of the problem. The paper describes the main issues arising in this integration: chromosome representation and evaluation, selection and replacement strategies, and the design of genetic operators. We also present a parallel execution model for a distributed memory architecture of the previous integration. We have adopted a global parallelization approach that preserves the properties, behavior, and fundamentals of the sequential algorithm. Linear speedup is achieved since genetic operators are coarse grained as they perform a search in a discrete space carrying out arc consistency. The implementation has been tested on a GRAY T3E multiprocessor using a complex constrained optimization problem.

Parallel Execution Models for Constraint Programming over Finite Domains

Many problems from artificial intelligence can be described as constraint satisfaction problems over finite domains (CSP(FD)), that is, a solution is an assignment of a value from a finite domain to each problem variable such that a set of constraints is satisfied. Arc-consistency algorithms remove inconsistent values from the set of values that can be assigned to a variable (its domain), thus reducing the search space. We have developed two parallelisation models of arc-consistency to be run on MIMD multiprocessors. Two different policies, static and dynamic, to schedule the execution of constraints have been tested. In the static scheduling policy, the set of constraints is divided into N partitions, which are executed in parallel on N processors. We discuss an important factor affecting performance, the criterion to establish the partition in order to balance the run-time workload. In the dynamic scheduling policy, any processor can execute any constraint, improving the workload balance. However, a coordination mechanism is required to ensure a sound order in the execution of constraints. Both parallelisation models have been implemented on a CRAY T3E multiprocessor with up to thirty four processors. Empirical results on speedup and behaviour of both models are reported and discussed.

Thinking precedes action: using software engineering for the development of a terminology database to improve access to biomedical documentation

Relational databases have been used to represent lexical knowledge since the days of machine-readable dictionaries. However, although software engineering provides a methodological framework for the construction of databases, most developing efforts focus on content, implementation and time-saving issues, and forget about the software engineering aspects of software and database construction. We have defined a methodology for the development of lexical resources that covers this and other aspects, by following a sound software engineering approach to formally represent knowledge. Nonetheless, the conceptual model from which it departs has some major limitations that need to be overcome. Based on a short analysis of common problems in existing lexical resources, we present an upgraded conceptual model as a first step towards the methodological development of a hierarchically organized concept-based terminology database, to improve the access to medical information as part of the SINAMED and ISIS projects.

Conceptual design for domain and task specific ontology-based linguistic resources

Regardless of the knowledge representation schema chosen to implement a linguistic resource, conceptual design is an important step in its development However, it is normally put aside by developing efforts as they focus on content, implementation and time-saving issues rather than on the software engineering aspects of the construction of linguistic resources Based on an analysis of common problems found in linguistic resources, we present a reusable conceptual model which incorporates elements that give ontology developers the possibility to establish formal semantic descriptions for concepts and relations, and thus avoiding the aforementioned common problems The model represents a step forward in our efforts to define a complete methodology for the design and implementation of ontology-based linguistic resources using relational databases and a sound software engineering approach for knowledge representation.

Parallel Evolutionary Optimisation with Constraint Propagation

This paper describes a parallel model for a distributed memory architecture of a non traditional evolutionary computation method, which integrates constraint propagation and evolution programs. This integration provides a problem-independent optimisation strategy for large scale constrained combinatorial problems over finite integer domains. We have adopted a global parallelisation approach which preserves the properties, behaviour, and theoretical studies of the sequential algorithm. Moreover, high speedup is achieved since genetic operators are coarsegrained, as they perform a search in a discrete space carrying out constraint propagation. A global parallelisation implies a single population but, as we focus on distributed memory architectures, the single virtual population is physically distributed among the processors. Selection and mating consider all the individuals in the population, but the application of genetic operators is performed in parallel. The implementation of the model has been tested on a CRAY T3E multiprocessor using two complex constrained optimisation problems. Experiments have proved the efficiency of this approach since linear speedups have been obtained.

Some Software Engineering and Knowledge Representation Principles for the Design and Construction of Ontology-Based Linguistic Resources

Nowadays, ontologies have become a common component of natural language processing systems as well as of many others. Nevertheless, in spite of the existing explosion of ontological resources, they have been developed without a rigorous methodological approach. Furthermore the usage of semantic relations between concepts has been so subjective that has led to errors affecting the coherence of these ontologies. In this article, we introduce a methodology for the construction of ontological linguistic resources based on software engineering, and develop the first of its 3 stages, introducing notions of knowledge organization and control of semantic relations for the gradual and systematic development of well-designed and coherent ontologies.

Bringing Forward Semantic Relations: Issues and Proposals

Semantic relations are an important element in the construction of ontologies and models of problem domains. Nevertheless, they remain fuzzy or under-specified. This is a pervasive problem in software engineering and artificial intelligence. Thus, we find semantic links that can have multiple interpretations in wide-coverage ontologies, semantic data models with abstractions that are not enough to capture the relation richness of problem domains, and improperly structured taxonomies. However, if relations are provided with precise semantics, some of these problems can be avoided, and meaningful operations can be performed on them. In this paper we present some insightful issues about the modeling, representation and usage of relations including the available taxonomy structuring methodologies as well as the initiatives aiming to provide relations with precise semantics. Moreover, we explain and propose the control of relations as a key issue for the coherent construction of ontologies.

Methodologically designing a hierarchically organized concept-based terminology database to improve access to biomedical documentation

Relational databases have been used to represent lexical knowledge since the days of machine-readable dictionaries However, although software engineering provides a methodological framework for the construction of databases, most developing efforts focus on content, implementation and time-saving issues, and forget about the software engineering aspects of database construction We have defined a methodology for the development of lexical resources that covers this and other aspects, by following a sound software engineering approach to formally represent knowledge Nonetheless, the conceptual model from which it departs has some major limitations that need to be overcome Based on a short analysis of common problems in existing lexical resources, we present an upgraded conceptual model as a first step towards the methodological development of a hierarchically organized concept-based terminology database, to improve the access to medical information as part of the SINAMED and ISIS projects.

Improving the Language Mastery through Responsive Environments

This paper is devoted to a way of intending to improve the language mastery of students in every subject and in a broad range of education levels by using computer-based tools involving linguistic resources. Lexicon concepts useful for language learning are involved in dictionaries and other kinds of linguistic resources. These concepts (vocabulary, meanings, semantic categories, semantic relationships, and taxonomy) are pointed out. Claim is made for using new environments and computer-human interfaces based on these concepts which define the pedagogical goals. Envisioning the foreseen collaborative instruction tasks, claim is made for the constructionist model of learning. Authoring and consulting electronic linguistic resources are the main tasks to reach the defined goals. The conceptual model appropriate for developing any software system taking into account all these principles is discussed, and the reached entity-relationship model is presented. We further present the developed tools for creating bilingual linguistic resources, which could allow reaching the foreseen learning goals. The authoring tool supports consistency of the intended semantics of the lexicon and enables the detection of omissions and inconsistencies, and the user tool is a graphical user interface for queries. These tools can be advantageously used in language teaching.

Parallel Implementation of Constraint Solving

Many problems from artificial intelligence can be described as constraint satisfaction problems over finite domains (CSP(FD)), that is, a solution is an assignment of a value to each problem variable such that a set of constraints is satisfied. Arc-consistency algorithms remove inconsistent values from the set of values that can be assigned to a variable (its domain), thus reducing the search space. We have developed a parallelisation scheme of arc-consistency to be run on MIMD multiprocessor. The set of constraints is divided into N partitions, which are executed in parallel on N processors. The parallelisation scheme has been implemented on a CRAY T3E multiprocessor with up to thirty-four processors. Empirical results on speedup and behaviour are reported and discussed.