Open AccessMULTIAGENT SYSTEM OF FORECASTING OF CHAOTIC PROCESSES ON THE BASIS OF STATISTICAL EXTRAPOLATORS
Author(s) -
Александр Азерович Мусаев,
Andrey V. Gaikov
Publication year - 2021
Publication title -
izvestiâ sankt-peterburgskogo gosudarstvennogo tehnologičeskogo instituta (tehničeskogo universiteta)
Language(s) - English
Resource type - Journals
ISSN - 1998-9849
DOI - 10.36807/1998-9849-2020-56-82-65-71
Subject(s) - chaotic , computer science , basis (linear algebra) , terminology , component (thermodynamics) , state (computer science) , artificial intelligence , data mining , mathematics , algorithm , linguistics , philosophy , physics , geometry , thermodynamics
The problem of the of a non-stationary system state predicting is considered. The decision based on the joint processing of the results obtained by a group of independent statistical extrapolators. In the terminology of multiagent systems, each extrapolator is an intelligent agent. The quality of the agent solutions is evaluated on retrospective data and is used as weight characteristic in the problem of a terminal solution estimation. The specificity of non-stationary processes with a chaotic system component leads to the empiricca version of the forecast generation algorithm