APPLICATION OF SMOOTHING METHODS FOR FORECASTING PRODUCTION VOLUME

Abstract

The object of the research is the enterprise LLC TPK «Terra» – the leading domestic manufacturer of high-quality protective coatings for industrial use for corrosion protection of metal structures and reinforced concrete structures. Now the company is open to planning future orders, the appropriate use of capacity and the expansion of production capacity. The paper conducted a system analysis of this enterprise by developing a functional model of processes and its decomposition with decomposition to the second level of detail. As a result, the task of forecasting the volume of release of anti-corrosion coatings, preventing the destruction of metal structures and mechanisms. This task is relevant both for the enterprise itself and for meeting the demand in the sales market. The most effective mathematical models that can be used to predict the development of production processes are models based on time series. One of the most common methods for predicting the performance of such series are smoothing methods that are used to reduce the effect of random fluctuations. The problem is solved by the methods of exponential and adaptive smoothing. To check the adequacy of the models obtained, the test of the series (determination of randomness of deviations from the trend), the criterion of the peaks (checking the equality of the mathematical expectation), the R/S criterion (determining the correspondence of the distribution of the residual component to the normal law), the Durbin–Watson criterion (determining the independence of the residual сomponents). The analysis of each prediction model obtained was carried out, and the quality of forecasts was also assessed. Conclusions are made regarding the further manufacture of products based on predicted values. The proposed model has a practical orientation and can be used in tasks related to forecasting in the conditions of industrial enterprises.