Evaluation of Control Priorities Based on a Matrix of Paired Comparisons

Abstract

The article contains a description of the model and approach for determining the weights of controls (other objects) based on a matrix of pairwise comparisons. The initial data in the matrix of paired comparisons implicitly specify interval estimates for comparison of controls. Interval estimates can be constructed relatively easily using Harrington's fuzzy metric scale. A heuristic method for determining the priorities of controls is proposed, which seeks to obtain weights (priorities) of controls that deviate minimally from the boundary values of the weight intervals.The latter are easy to find using the system of relations, on the basis of which the matrix of pairwise comparisons of controls is built. To assess the statistical adequacy of the found control priorities, the determination coefficient is used, the value of which at the level of 80 % and above is considered practically acceptable for applications. The critical value of the coefficient of determination can be determined, for example, using the Fisher criterion. If the conditions of statistical adequacy are not met, the initial data in the matrix of paired comparisons should be revised. An example is provided to illustrate the method described in the article. The main advantages of the approach described in the article are the flexibility of the choice of interval estimates for pairwise comparison of controls, the relative simplicity of computational calculations, and the ability to control the calculation results using statistical criteria. The described method develops the authors' approach to the adoption of anti-crisis management by adapting management priorities to the conditions of the production system and its environment.