Consistency-based and abductive diagnoses as generalised stable models

If realistic systems are to be successfully modelled and efficiently diagnosed using model-based techniques, a more expressive language than classical logic is required. In this paper, we present a definition of diagnosis which allows the use of a nonmonotonic construct, negation as failure, in the modelling language. This definition is based on thegeneralised stable model semantics of abduction. Furthermore, we argue that, if negation as failure is permitted in the modelling language, the distinction between abductive and consistency-based diagnosis is no longer clear. Our definition allows both forms of diagnosis to be expressed in a single framework. It also allows a single interference procedure to perform abductive or consistency-based diagnosis, as appropriate.This paper is an extended and revised version of ref. [29].     

Testable heuristics for the 0–1 knapsack problem

Two kinds of heuristics, fixed time and cut time, are proposed in order to use the running time available in solving 0–1 knapsack problems profitably.Translated from Kibernetika, No. 2, pp. 44–52, March–April, 1991.     

The new intra-urban dynamics: Suburbanisation and functional specialisation in French cities

This article examines the relationships between the employment suburbanisation from central cities towards their suburbs, and the process of intra-urban specialisation that occurred simultaneously in the fifty largest French metropolitan areas. A methodology is proposed to identify urban subcentres and to analyse the effects of the intra-urban specialisations on suburbanisation patterns. We conclude that the specialisation of both subcentres and central cities has a significant effect on suburbanisation rates. Lastly, an intra-metropolitan shift/share analysis provides additional insights into the employment dynamics of central cities and suburbs during the last twenty years. Received: 25 July 2000 / Accepted: 29 May 2001     

Effect of temperature and cyclic heating on the structure and strength of glass textolite

The article established the regularities of change of the physicomechanical characteristics and strength of glass textolites in single and repeated thermal loading. It shows that under multiple loading the intense loss of strength of the investigated materials is due to increased porosity on account of the large number of microcracks appearing in the structure. A correlation is obtained between the elastic and strength characteristics and porosity in dependence on the temperature of multiple heating which in the first approximation is described by a polynomial of second degree with respect to temperature, multiplied by the reciprocal of porosity.Translated from Problemy Prochnosti, No. 6, pp. 53–58, June, 1991.     

Doped silicon and NIS junctions for bolometer applications

We discuss the performance, of a normal metal hot electron bolometer (NHEB) that we have measured at 0.3 K. We found that the noise equivalent power was limited by the amplifier noise. To improve the NHEB power response and to make it more robust and reliable we propose to substitute the normal metal with heavily doped silicon. The heavily doped silicon behaves like a metal with lower carrier concentration and has a smaller electron–phonon thermal coupling. We have fabricated superconductor-doped silicon-superconductor contacts (S-Sm-S) and we have used them as thermometers and coolers.     

Converting harmonic oscillations into chaos

A new simple model of a system with chaotic dynamics, based on the equations of bistable systems, is considered. The possibility of converting harmonic signals into chaotic oscillations, which represent intermittent irregular and switching quasi-regular motions, is demonstrated by numerical methods. The mechanism of chaotization is analyzed using the results of numerical calculations.     

An on-line learning algorithm for the orthogonal weight estimation of MLP

We propose an on-line learning algorithm for Multi Layered Perceptrons (MLP) with an Orthogonal Weight Estimator (OWE) architecture. Such an architecture allows to dynamically and efficiently estimate the weights of a MLP in context dependent behaviour problems. The proposed learning algorithm attempts to solve the problem of time-consuming in the learning phase encountered to train these weight estimators.