Swarm intelligence in the problems of detecting anomalies and states of natural-technical systems and objects

An approach to solving the problem of detecting and classifying anomalies and states of natural-technical systems and objects using swarm intelligence methods is considered. The main directions of development of the proposed approach include ant algorithms, bee swarm algorithms, and the particle swarm method. The structure of the swarm intelligence system of decision support based on collective preference rules is proposed. The application of the proposed approach makes it possible to optimize the processes of processing, analysis, integration of heterogeneous data, to increase the sensitivity, reliability and efficiency of decisions made.

Download Full-text

Decision Support for Performance Arts Using Support Vector Regression With Genetic and Particle Swarm Algorithms

Volume 2A: 41st Design Automation Conference ◽

10.1115/detc2015-47059 ◽

2015 ◽

Author(s):

Meifang Li ◽

Mian Li

Keyword(s):

Decision Support ◽

Support Vector Regression ◽

Performing Arts ◽

Influence Factor ◽

Particle Swarm ◽

Consumer Products ◽

Support Vector ◽

Strategic Decisions ◽

Single Output ◽

Swarm Algorithms

Different from typical mechanical products, tickets for movies and performing arts can be considered as a special type of consumer products. Compared to widely known box-office receipts prediction with single-output in movie industry, estimating the market share and price for performing arts is still a challenging problem due to high dimensional datasets yet limited number of samples. This paper describes a data-driven decision support system to help arts managers make strategic decisions, especially on session-determination and price-setting, considering price discrimination and prediction on the corresponding sales volume. Eight different attributes from the database, with multiple labels in each attribute, are used to accurately and comprehensively represent and classify the characteristics of performing arts in each genre. A web-based influence factor is also defined to quantify the popularity and publicity of performing arts. For this multi-input and multi-output problem, support vector regression (SVR) is employed and its optimal parameters are determined using genetic algorithm (GA) and particle swarm optimization (PSO) respectively. Price utility axiom with the law of demand is applied to maximize the receipts. Compared to artificial neural networks (ANN), those two optimization based SVR methods perform much better, in terms of effectiveness and reliability.

Download Full-text

Creation and Implementation of Information Reserves in Big Technical Systems Maintenance

Proceedings of Telecommunication Universities ◽

10.31854/1813-324x-2020-6-4-104-110 ◽

2020 ◽

Vol 6 (4) ◽

pp. 104-110

Author(s):

A. Shestakov ◽

K. Biliatdinov

Keyword(s):

Decision Support ◽

Complex Systems ◽

Experimental Tests ◽

System Engineering ◽

Heterogeneous Data ◽

Industrial Facilities ◽

Proactive Management ◽

Systems Maintenance ◽

Technical Systems ◽

New System

The introduction and use of modern methods of proactive management of complex systems in decision support systems based on the collection and processing of heterogeneous data about the system and the generated precedents of events and corrective actions requires the search for new system-technical solutions for the data collection, storage and preliminary analysis, taking into account the control conditions and the characteristics of complex systems. A special feature of complex systems, such as large technical systems of industrial facilities, informatization and communication, transport, etc., is the combination of improvement stages (modernization of components) and development of systems (introduction of fundamentally new system engineering solutions) during their regular operation. In this regard, it is quite relevant to study the model for maintaining large technical systems and ways to create databases based on information reserves for decision support of proactive management. For the first time, system engineering solutions for the implementation of the original model and method based on a set of computer programs and the results of experimental tests are presented.

Download Full-text