APPROACHES TO RISK FORMALIZATIONS FOR EVOLUTIONARY SYSTEMS MANAGING

Risk is an interdisciplinary concept, and in some cases, it can function as a category. It is associated with human activity and reflects the impact of uncertainty on the results of activities. It can have both quantum (risk as an event) and wave (risk as a process) characteristics. To use the concept of risk constructively in management tasks, it is necessary to formalise it. Risk can be formalised in terms of three fundamental dimensions: possibility (frequency), consequences (damage) and goal achievement (quality). The pairwise combination of these properties in the formalisation of risk leads to nine formal measures: Bayesian risk (a posteriori estimation of error); risk as a measure of stability in phase space; risk as a measure of inconsistency of expert judgments; risk as a combination of probability and damage of an event; relative risk; risk as an anti-potential of development; risk as a measure of difference of states; risk as a consequence of playing with nature; risk as the difficulty of achieving a goal. A management model of an evolving system is considered, which makes it possible to combine the obtained formal risk measures in a paradigm, which provides the possibility of long-term forecasting of the behaviour of the system under study and modelling the reaction of this system to a given change in the parameters describing it and the environment external to it. As a basis for the complex joint use of the following basic ideas are proposed: a method of tracking the dynamics of the selected indicator in the phase space of states; tracking not the indicator itself, but the centre of quasi-attractors of its trajectories; construction of the area of permissible values of the indicator (the centre of quasi-attractors) in this space; if there is a target value, tracking the risk of not achieving the target and determining the frequency of motion control; tracking the causes of variability of the system under study, identifying special causes of variability; predicting the state of the system in the future by the movement of the centre of quasi-attractors. The joint application of these ideas is a method of stress testing the system under consideration to significant external and internal changes, which makes it possible to find the limits of its stability and work out scenarios for managing the system in emergency and crisis situations.