• A
  • A
  • A
  • ABC
  • ABC
  • ABC
  • А
  • А
  • А
  • А
  • А
Regular version of the site

Article

Адаптация методов семейства ELECTRE для включения в Экспертную систему поддержки принятия решений

Бизнес-информатика. 2015. № 2 (32). С. 69-78.
Кравченко Т. К., Дружаев А. А.

Analytical justification of solutions using a decision support system (DSS) significantly increases the quality of decisions. The existing DSS generally employs 1 or 2 methods of decision-making. It does not always lead to the desired results, as each method is based on certain assumptions and is not universal. The maximum effect can be achieved only insofar a set of decision-making methods is included into the knowledge base of the DSS. The only system that meets these requirements is the Expert Decision Support System (EDSS) developed under supervision of the author. Currently the EDSS includes about 50 decision-making methods. The expansion of the EDSS knowledge base by including new methods will allow for choosing the most suitable method for solving each decision-making task. Enhancing the Decision Table model underlying the system knowledge base allows you to develop the EDSS without complete reworking of the system code. The system knowledge base contains decision rules built on the principle of “if... then...” (if certain conditions of decision making exist, then a definite method of decision-making should be employed). To expand the EDSS knowledge base, ELECTRE collection methods were selected. Their key feature consists in not using the convolution operation of evaluation of the alternatives specified in different scales on individual criteria. This was the reason for selecting the methods of this family. In the article, the algorithms of these methods are adapted for their inclusion in the EDSS. An algorithm for obtaining a criterion-alternative matrix is proposed. It serves as input information for the ELECTRE family methods in cases where there is no objective information for its formation. The results of the study can be used to develop the EDSS, allowing analytical justification of solutions using methods that have not previously been used in the system.