Open AccessStability model of agent-based situational control system
Author(s) -
Oleksii Kovalenko,
Vladimir Kosolapov
Publication year - 2020
Publication title -
matematičeskie mašiny i sistemy
Language(s) - English
Resource type - Journals
ISSN - 1028-9763
DOI - 10.34121/1028-9763-2020-3-93-104
Subject(s) - context (archaeology) , computer science , process (computing) , stability (learning theory) , complex system , situational ethics , dissipative system , artificial intelligence , machine learning , paleontology , physics , quantum mechanics , political science , law , biology , operating system
While managing complex systems, it is advisable to use mathematical models adequate to real systems, which can be used for generalized model analysis of different solutions in systems of situational man-agement (SSM). Such mathematical models are an important component of the SSM in the process of supporting the adoption of important strategic and operational decisions at various levels of government. Implementation of SSM in the form of a multi-agent system, due to its characteristics, is an adequate approach to solving the problems of situational management (SM). According to the context of SM, the behavior of the SSM is described as the dynamics of movement from a certain point in the phase space, that corresponds to some state of the managed system under the influence of the ensemble of SSM agents. During the operation of the SSM, its agents use knowledge that corresponds to the context of the situation. The agent's knowledge is a fragment of the field of knowledge on the target problem of SM. Knowledge of the problem area of the SM is a key element of the SM model. The convergent agents’ ensemble of SSM is characterized by a certain level of intelligence, which is represented as an entropic force that uses the free energy of the dissipative system to maintain its stability. An agent-oriented approach to the study of the stability of a dynamic stochastic system in the process of situa-tional management as a target project activity is proposed. Within the scope of the proposed approach, the stability model of agent SSM as a dynamic stochastic system is considered using the Lyapunov sta-bility criterion in the form of a system of ordinary differential equations. Support of modeling functions by agents of agent-oriented system allows to form adequate behavior in the process of situational man-agement in the conditions of changes in the environment. Developed models for integrating behavioral and coordination aspects of knowledge-based agents can be used in the development of situational man-agement systems and technologies.