Hide
Раскрыть

ISSN 2587-814X (print),
ISSN 2587-8158 (online)

Russian version: ISSN 1998-0663 (print),
ISSN 2587-8166 (online)

V. Belov, J. Koritchneva

Multivariate ABC-classification. Criteria of quality and initial algorithms

2012. No. 1(19). P. 9–16 [issue contents]

Vladimir Belov – Professor, Department of Computational and Applied Mathematics, Faculty of Computer Technologies, Ryazan State Radio Engineering University (RSREU).
Address: 59/1, Ryazan, Gagarin str., 390005, Russian Federation.
E-mail: compvv@mail.ryazan.ru

Julia Koritchneva – Post-graduate student, Department of Computational and Applied Mathematics, Faculty of Computer Technologies, Ryazan State Radio Engineering University (RSREU).
Address: 59/1, Ryazan, Gagarin str., 390005, Russian Federation.
E-mail: Koritchneva@mail.ru

The classical ABC method is based on splitting the whole set of the materials in use into three unequal groups depending on the value of the defining indicator (parameter, criterion). However, the practical implementation of integral parameters showed that the results of the multidimensional classification are often hard to predict and do not correspond with the intuitive solutions of the experts. Thus, it is necessary to develop alternative methods of multidimensional ABC- classification, including formulating a parameter for assessing quality of multidimensional classifications. The aim of the article is to formulate the parameters identifying quality and the methods of a multidimensional ABC- classification for data managing, which could be an alternative to the integral criterion method.

The article offers alternative forms of presentation of the results ABC-classification – in the form of classification tuples and vectors. Four alternative parameters, describing the quality spatial (multidimensional) ABC-classification, are proposed, which reflect the similarity of the spatial vectors and tuples and aggregated private scalar ABC-classifications for different cases of the same problem and the importance of scalar criteria used to characterize the accounting elements particular subject area. The article proposes four algorithms of spatial ABC-classification, called canonical: most suitable on the basis of the proposed multidimensional measures of the quality of the grouping for cases of the same and different the problem importance of private scalar criteria.

The results presented here can be used as a methodological platform for implementing the means of reducing the information space in the logistics practices in order to improve  inventory management through the expedient and reasonable distribution of efforts in various areas of control of the situation and the development of control measures.

The solution for multidimensional ABC-classification through the proposed canonical algorithms, comparison of the achieved results and comparison with the result of ABC-classification based on the convolution of criteria implemented by private credentials inventory of high-tech manufacturing enterprise engaged in production of small-scale and technological piece of equipment having a well-oiled mechanism of delivery materials, components and related products in the shop just in time.

Citation: Belov V. V., Koritchneva J. L. (2012) Mnogomernaia AVS-klassifikatciia. Kriterii kachestva i kanonicheskie algoritmy. [Multivariate ABC-classification. Criteria of quality and initial algorithms] Biznes-informatika, 1(19), pp. 9-16 (in Russian)
BiBTeX
RIS
 
 
Rambler's Top100 rss