Search

This paper deals with the key optimization task that has to be solved when improving the performance of many chemical processes--optimization of the catalysts used in the reaction via the optimization of its composition and preparation. A novel approach is presented that allows for the preservation of the advantages of genetic algorithms developed specifically for the optimization of catalytic materials but avoids the disadvantageous necessity to reimplement the algorithm when the scope of the optimized materials changes. Its main idea is to automatically generate problem-tailored implementations from requirements concerning the materials with a program generator. For the specification of such requirements, a formal description language, called catalyst description language, has been developed.

This article addresses the topic of extracting logical rules from data by means of artificial neural networks. The approach based on piecewise linear neural networks is revisited, which has already been used for the extraction of Boolean rules in the past, and it is shown that this approach can be important also for the extraction of fuzzy rules. Two important theoretical properties of piecewise-linear neural networks are proved, allowing an elaboration of the basic ideas of the approach into several variants of an algorithm for the extraction of Boolean rules. That algorithm has already been used in two real-world applications. Finally, a connection to the extraction of rules of the Łukasiewicz logic is established, relying on recent results about rational McNaughton functions. Based on one of the constructive proofs of the McNaughton theorem, an algorithm is formulated that in principle allows extracting a particular kind of formulas of the Łukasiewicz predicate logic from piecewise-linear neural networks trained with rational data.

Exploration of a complex catalyst system using Genetic Algorithm methods and combinatorial experimentation efficiently removes noncontributing elements and generates data that can be used to model the remaining system. In particular the combined methods effectively navigate and optimize systems with highly nonlinear dependencies (3-way and higher interactions).

Data mining has been recently experiencing a boom of interest from researchers and software producers. In medicine, however, its applications are still rather rare. In this paper, we argue that this is primarily due to the requirements of reproducibility of results and diversity of available data mining tools, both of which are crucially important for medical research. We propose to tackle the diversity requirement by means of distributed object technologies. The results presented here rely on our experience with medical data mining using the method GUHA and with further developing that method.

The structure of healthcare systems in developed countries is changing to 'shared care', enforced by economic constraints and caused by a change in the basic conditions of care. That development results in co-operative health information systems across the boundaries of organisational, technological, and policy domains. Increasingly, these distributed and, as far as their domains are concerned, heterogeneous systems are based on middleware approaches, such as CORBA. Regarding the sensitivity of personal and medical data, such open, distributed, and heterogeneous health information systems require a high level of data protection and data security, both with respect to patient information and with respect to users. This paper, relying on experience gained through our activities in CORBAmed, describes the possibilities the CORBA middleware provides to achieve application and communication security. On the background of the overall CORBA architecture, it outlines the different security services previewed in the adopted CORBA specifications which are discussed in the context of the security requirements of healthcare information systems. Security services required in the healthcare domain but not available at the moment are mentioned. A solution is proposed, which on the one hand allows to make use of the available CORBA security services and additional ones, on the other hand remains open to other middleware approaches, such as DHE or HL7.

The development of healthcare systems in accordance to the "Shared Care" paradigm results in co-operative health information systems across the boundaries of organisational, technological, and policy domains. Increasingly, these distributed and heterogeneous systems are based on middleware approaches, such as CORBA. Regarding the sensitivity of personal and medical data, such open, distributed, and heterogeneous health information systems demand a high level of data protection and data security, both with respect to patient information and with respect to users. The security concepts and measures available and additionally needed in health information systems based on CORBA architecture are described in this paper. The proposed security solution is also open to other middleware approaches, such as DHE or HL7.

To meet the challenges for efficient and high-level quality, health care systems must implement the "Shared Care" paradigm of distributed co-operating systems. To this end, both the newly developed and legacy applications must be fully integrated into the care process. These requirements can be fulfilled by information systems based on middleware concepts. In the paper, the middleware approaches HL7, DHE, and CORBA are described. The relevance of those approaches to the healthcare domain is documented. The description presented here is complemented through two other papers in this volume, concentrating on the evaluation of the approaches, and on their security threats and solutions.

Increasingly, distributed, interoperable healthcare information systems, which meet the shared care paradigm, work across the boundaries of policy, organisational, and technological domains and are based on middleware concepts. Especially in healthcare with its sensitive personal and medical data, such systems require advanced data security measures. In the paper, a common object-oriented security model for middleware systems and advertisements for implementation are proposed, corresponding the security requirement of both the user and the application environment.

Regarding the need of change of both healthcare systems' structure and supporting information systems' architecture to the "Shared Care" paradigm, appropriate health information systems have to be distributed and interoperable, and will usually be based on middleware concepts. Structure and function of information systems meeting such requirements have been analysed in general. To describe, evaluate and combine distributed systems, a generic component model has been used. As a result, CORBA, DHE, and HL7 have been assessed systematically. A way to a common and harmonised middleware approach has been proposed.