ECOLOGICAL RESILIENCE ASSESSMENT FEATURES OF NATURAL-INDUSTRIAL SYSTEMS BY INDEX-BASED METHODS USING SOFT COMPUTING

Background. The study is aimed at developing approaches and methods for information support of multi-criteria decision-making in management problem-solving of natural-industrial system entities in the Arctic region of Russia in order to ensure their environmental safety and resilient operation under the influence of negative natural and man-made factors. Regional peculiarities and specific character of economic development condition on the vulnerability of entities localized in the territory of a given region in terms of the emergence of adverse triggering events of various genesis, necessitating the development of effective measures and policy decisions that enhance the reliability and safety of existing regional infrastructure entities. To make informed decisions and a reasonable choice of measures to counteract current threats when managing safety and resilience of ecosystems, the evaluation and analysis of alternative decisions based on multiple criteria should be carried out. Materials and methods. The unified methodological framework of our study includes the fundamentals of safety sciences, risk management methodology, and the stateof- the-art resilience concept of complex systems, as currently developing research area. The proposed approach to a multilevel integral estimation of environmental safety of territories is based on the combined use of index-based methods for quantitative and qualitative analysis of heterogeneous indicators of ecosystem resilience and hierarchical aggregation of the assessments. Results and conclusions. A modified index-based method, expanding the capabilities of existing methodologies for calculating the integral index of environmental conditions by incorporating qualitative indicators and situational factors, taking into account their unequal significance and antagonism, as well as maintaining the balance of indicators, has been developed. The method is based on the automatic transformation of estimates into membership functions and fuzzy convolution of the normalized indicators across levels of the organizational hierarchy. Proposed method allows for a quantitative assessment of the exposure degree of natural-industrial system entities on the state of ecosystem resilience under incompleteness and uncertainty of the raw data, and can be used to prepare and implement a set of preventive and reactive measures to improve the environmental safety of the territory when providing control and impact monitoring of these entities.