Verification of Scenario Models: Technology Analysis

Abstract. This paper is devoted to methodological problems of increasing the efficiency of the scenario approach in the processes of planning and managing the development of social and economic systems under uncertainty. Ways to improve the adequacy of simulation models based on the mathematical language of functional digraphs, used to form and analyze alternative scenarios for the evolution of problem situations, are presented. Mechanisms for analyzing a subject area and designing a problem situation passport are considered. The resulting part of this passport contains a problem statement and an initial scenario model based on coordinated expert assessments. To improve the adequacy of the basic simulation model yielded by the program-analytical complex of scenario modeling, the idea is to use verification testing mechanisms. They are intended to diagnose and isolate errors in determining the set of model factors and the properties of their relations, as well as to correct the errors by adjusting the model structure. Based on the concept of a typical verification testing scheme introduced below, methods for assessing the efficiency of verification procedures (the probability of no errors and the testing time) are developed. To improve the quality of scenarios generated and evolution models of problem situations, a scenario-event approach is proposed. This approach implements a transition from a real-valued phase space reflecting changes in the values of factors of a scenario simulation model to an event space. The approach proposed expands the capabilities to analyze the processes and problem situations in qualitative terms, assess the correct realization of causal inter-factor and inter-model relations, and analyze the consistency of the modeling results on a set of identifiable events. Some examples are provided to illustrate the practical use of the above approaches, methods, and technologies in solving several application-oriented problems of managing the sustainable and safe development of complex socio-economic systems.

Keywords: scenario analysis, control, uncertainty, simulation modeling, model verification, testing.