Optimizing Model of a Machine-Building Enterprise Output under Demand Uncertainty

Abstract

The article develops anoptimizing model for a machine-building enterpriseproduction plan, taking into account a combination of factors such as financial capabilities and production capacity of the enterprise, limited resources, as well as demand uncertainty of long-range planning for a period of more than three months. Product features of a machine-building plant is taken into account that produces equipment and spare parts for oil-extraction enterprises, as well as the peculiarities of its operations in the market, due to the capabilityof sales forecasting for 2-3-month periods ahead based on actual demand under tender agreements. And this will be the minimum plan that the production company needs to follow. It is proposed to solve the demand uncertainty by means of adaptive forecasting methods based on statistical data available from thesales enterprise indicatorsof the past periods. It is econometric models ARMA and ARIMA that are designed to analyze time series that do not contain a seasonal component, since the demand for the products of a machine-building enterprise has no seasonal fluctuations. The ARMA model uses a combination of past values and white noise to predict future values and assumes that time series data are stationary, which means their statistical properties will not change over time. The ARIMA model is an extension of ARMA for non-stationary time series and is used to build short-term forecasts in which the lead period ranges from one or several months to a year. The mathematical model is a nonlinear conditional optimization problem. To solve this problem, various methods are proposed, allowing determination the optimal production plan, the minimum loan amount and the resources needed to implement this plan.