Theory and practice of using fuzzy Petri nets for monitoring environmental risks

The results of theoretical analysis of fuzzy production systems for decision support in the context of geoinformation monitoring of environmental risks are presented. Along with classical approaches to structural and functional construction of geoinformation systems of GIS analysis and monitoring of environmental risks, a new approach to the organization of the process of analysis and monitoring of environmental risks in conditions of fuzziness is considered. The approach is based on applying modern theory and practice of risk analysis and monitoring in complex software projects. In the proposed approach, the process of analyzing and monitoring environmental risks in the GIS includes an algorithm consisting of the following steps: identification of environmental risks; a qualitative analysis of environmental risks; quantitative analysis of environmental risks; planning of response to environmental risks; monitoring and risk management. An important step of this algorithm is a qualitative analysis of environmental risks, performed by expert methods. As a result of this analysis, experts form a ranked register of environmental risks, which are grouped by category. The established register of environmental risks is the basis for the following stages of the algorithm for system analysis and monitoring of environmental risks. At the same time, the process of qualitative analysis of environmental risks is a very laborious process. To formalize this process, it is proposed to use modern methods to support the adoption of expert decisions in conditions of fuzziness. In this connection, the article considers the approach to presenting fuzzy product rules in a fuzzy production GIS designed for the analysis and monitoring of environmental risks. Existing fuzzy decision-making decision support systems by definition are designed to implement the process of fuzzy inference and serve as a conceptual basis for modern fuzzy logic. The achieved successes in the application of these systems for solving a wide class of control tasks and served as the basis for choosing a mathematical apparatus for formalizing the process of analysis and reducing environmental risks based on the use of models, methods and algorithms of fuzzy sets theory and fuzzy Petri nets. Within the framework of the proposed formalized approach to risk analysis, NSPs obtained as a result of the introduction of fuzziness in the initial marking and in the rules for triggering the transitions of the basic formalism of ordinary JVs are considered. The considered system analysis of the NSP allows us to concretize the proposed way of presenting the rules of fuzzy products in solving applied problems of fuzzy modeling and performing the process of approximate reasoning on environmental risks. Based on the use of modified NSPs, a pilot version of an automated decision support system for environmental risks can be developed. This system is essentially an expert system that reflects the fuzzy logic of the interrelation of input values - expert assessments of the state of the region's ecology and output values - the degrees of truth of possible environmental risks. The main functional modules of such an automated system should be considered: a base of rules for fuzzy products, a block of fuzzy logic inference, an interface module, a rule base modification module, modules for fuzzification and defuzzification. The user interface of the automated system should be based on the representation of all system objects and functions available to the user in the form of graphic components of the screen. To formalize the process of environmental risk analysis, the paper proposes a way to present fuzzy product rules in a fuzzy environmental risk analysis system based on the use of fuzzy Petri nets. A program implementation of a fuzzy production system for the analysis of environmental risks based on the use of modified fuzzy Petri nets was implemented. The software product has a sufficiently effective application for solving problems in the GIS of environmental monitoring. The modern theory and practice of GIS development for the analysis and monitoring of environmental risks in conditions of diverse initial geodata indicates the need to develop new effective approaches and algorithms for supporting decision-making on environmental risks in conditions of fuzziness. As a result of the software implementation of the algorithm for analyzing environmental risks using the mathematical apparatus of the NRS, an effective and easy-to-use software product has been developed that is intended for application in practice both independently and as part of industrial GIS.

Keywords

Authors

References