Automation of selecting the functional structure of a system for complex emergency simulation

Abstract

The paper formulates the problem of selecting the composition of computational and modeling services that form the functional structure of complex modeling systems (CMS) of emergencies and describes the algorithm for its solution. The main initial prerequisites for the formulation of the mentioned problem are: 1 - the need to use not one, but a set of models, with the implementation of a mechanism for the prompt selection and adjustment of specific models in each particular emergency situation to improve the quality of emergency forecasting; 2 - the feasibility of using a service-oriented architecture to create an CMS, within which the computational and modeling components of the system are implemented as web services with a certain set of values of their functioning quality indicators. The selection problem is posed as a multi-criteria one with three types of quality indicators - the cost of modeling (use of services), operativeness (duration of calculation operations) and the target quality indicator (accuracy of modeling). As a result of solving this problem according to the proposed algorithm, a compromise version of the functional structure of the CMS is determined, which minimizes the maximum relative weighted deviation from the optima by particular indicators of the quality of services. The demonstration example shows that the proposed approach provides a fundamental possibility of automating the task of selecting the CMS functional structure both directly when modeling a developing emergency, and in a scenario mode.