Europe

INTAS 96-0952

Field: Mathematics,Telecommunications, Information Technologies

Subfield: Computer Software

INTAS funding: 60000 ECU

Duration: 36 months, 1997-2000

Report Reference

Aristotle University (Thessaloniki, Greece), Metadynamics Ltd. (Georgiou I), Universidad Politecnica de Madrid (Spain), Institute for Informatics and Automation Problems (Yerevan, Armenia), Computing Center of Academy of Sciences of Russia (Moscow) (2000) Concurrent heuristics in data analysis and prediction.

Final Report here

Coordinator:

Professor Grigorios Tsagas

School of Technology

Division of Mathematics

Egnatia

GR - 54 006 Thessaloniki

Tel: +30 31 995973

Fax: +30 31 996155

E-mail: tsagas@vergina.eng.auth.gr

Partners:

Phokion 'Ion' Georgiou

Metadynamics Ltd.

UK - SW13 9QX London

Levon Aslanyan

Institute for Informatics and Automation Problems

Laboratory of Discrete Analysis and Modelling Technologies

AM - 375014 Yerevan

Juan Castellanos

Universidad Politecnica de Madrid

Facultad de Informatica

ES - 28660 Boadilla del Monte - Madrid

Vladimir Riazanov

Computing Centre

Dept. of Recognition Problems and Combinatorial Analysis

RU - 141700 Moscow

Concurrent heuristics in data analysis and prediction

Abstract by Levon Aslanyan

The main objective of this research is to address the technology development for modelling of broad range application problems based on competitive strategies or heuristic data analysis ranging from theoretical to practical aspects. More specifically, these activities will focus on the study of the sets of discrete optimization problems common for many applications as well as of pattern recognition and heuristic reasoning modelling of information technologies. The two of the so called critical social issues: Geophysical monitoring and Industrial information management basically will be addressed as applications.

Geophysical monitoring involves several issues, ranging from data acquisition, to data analysis and to the development of prediction criteria of anomaly or predictive events.

Industrial management problems in nowadays transitional period in politics and economy of FSU countries concern distributed integrated information structures for decision making with different planning strategies.

These and similar issues are present both in European countries and in NIS countries, and are of primary importance for the whole European community. The idea is to develop the adequate formalisms for these badly formalized problems rather than the use of incompatible technic and making the corresponding information systems and technologies freely and widely accessible via Internet.

In theory aimed to initiate and develop new mathematical approaches and related algorithms for heuristic data analysis and forecasting, based on formalization of logical consistency description of applicational problems and creation of metaprocedures above the sets of heuristic procedures lead to the reasoning. These overall structures consists of informational matrices, and algorithms of discrete optimization on permutations. They include the additional characteristics and properties of problems to be solved and provide the creation of appropriate metaprocedures for the given conditions of the considering problems and their acceptable solutions.

In practice aimed to set up the basis for a larger cooperation network on the issues of consultancy and formalization of special classes of critical and ill defined applicational problems. They addressed to the newly understudied research and application situation in science and industry based on complexity, incompleteness and contradictoriness of a priori empirical and expert data consistence. This restricts or even excludes the possibility of mathematical models of objects, situations, processes and phenomena construction with the only use of classical mathematical methods of heuristic data treatment. As well as new algorithmic solutions will be achieved and new approaches to the sets of heuristic procedures and models will be developed, the two basic and a number of specialized software systems will be necessary to create.

Papers, research reports and software systems will introduce the main research results of this project.