Parameter Identification of the Agent-Based Model for Managing a Regional Industrial Complex

Abstract

Current challenges facing the Russian economy require models to optimise industrial management processes at the regional level. Therefore, a three-level hierarchical agent-based model for minimax control of the regional industrial complex is considered in this research. The study aims to develop an approach to solving a parameter identification problem of the agent-based model for managing the industrial complex of Sverdlovsk oblast. To this end, the paper presents a theoretical justification for the implemented approach, a formalisation of the aforementioned problem, and an algorithm for constructing and selecting models to assess management system parameters. The method of linear regression analysis was applied to solve the identification problem. The proposed approach was tested using data on 28 types of industrial activity in Sverdlovsk oblast for 2005–2021. The phase vector of statistical identification models was defined by the following parameters: the average annual number of employees of enterprises; fixed assets; gross value added; volume of shipped goods, performed works and services; balanced financial results of enterprises; investment in fixed capital; costs of implementing and using digital technologies. The control vector was determined by the factors of attracting budgetary and other (external) funds. As a result, 125 high-quality models were built to solve parameter identification problems of the agent-based model used for managing the industrial complex of Sverdlovsk oblast. The obtained statistical models can be used to establish communication between agents, clarify their specificity, calculate and assess management performance. The proposed approach can be applied to predict the development of regional industrial complexes in accordance with planned control actions, as well as to calculate control actions necessary to achieve target parameters.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 5 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.