Methodical Aspects of Risk Analysis of Target Underachievement in Forming the Mid-Term Forecasts of Development of Structurally Complex Systems

Traditional models of forecasting market and economic behaviour (belonging to the class of purposeful structural-complex systems) are based in most cases on the analysis of the existing and retrospective balance of resource extraction by world exporters and resource consumption by importers, considering the development forecasts of their industrial production and power engineering. The availability of long-term back data makes it possible to use production functions or multi-factor regression models to make short-term forecasts. As uncertainty increases with the length of the forecast horizon, the accuracy of the forecasts decreases. The corridor of acceptable values in the forecasting model is therefore determined by the degree of variability of the backward time series used to build the model. The latter in turn depends on the growth (decline) of demand/consumption in the past. For modelling, the maximum discrepancies in the data are often smoothed out, which leads to a situation where the model makes a forecast on the basis of a time series that differs from the initial one, and therefore it is unable to predict the approaching crisis. The approach proposed by the authors is based on the actual (or declared) value of the maximum and minimum variability of the forecast indicators, which defines a forecast corridor in each time interval set for reaching the target state of the indicators. Thus, it is not a point value that is assessed, but its achievability within the corridor of admissible values and the existing quality of the developing system.