Technology for risk assessment at product lifecycle stages using fuzzy logic

The problem of risk assessment at the stages of the product life cycle using both qualitative and quantitative approaches is investigated, and a generalized algorithm for selecting a fuzzy risk assessment model with different input data and system requirements is proposed for the effective use of statistical information and expert assessments. The "risk-based approach" allows to reduce the cost of correcting possible errors in the future and reduce the uncertainty when performing subsequent actions. It is noted that the results of SWOT analysis, as a rule, are of a qualitative descriptive nature, and do not contain specific recommendations. The provisions of modern standards on risk analysis are analyzed and the classification of risk analysis methods is given in accordance with the provisions of the national standard GOST R 58771-2019 "Risk management. Technologies for risk assessment", in which the key is the concept of uncertainty, estimated using different scales of gradation of risk damage and probability of its occurrence. An approach based on fuzzy logic and a hybrid fuzzy neural network model is proposed, which allows to present the used criteria in a con-venient form and implement a logical conclusion using simple and visual production rules. At the same time, the effectiveness and accuracy of the developed risk assessment system based on fuzzy logic is mainly determined by the quality of expert information and the consistency of the methods used to obtain it. To improve the accuracy of the results, it is proposed to use collective expert estimates with subsequent analysis of the consistency of the obtained expert estimates by determining the coefficients of variation, rank correlation, concordation, and so on. A generalized algorithm of expert assessment is presented, which is recommended to follow when developing expert systems for risk analysis. Various models of fuzzy inference (Mamdani, Takagi-Sugeno, hybrid neuro-fuzzy inference) are considered. An algorithm for constructing a fuzzy risk analysis system based on an effective method for obtaining expert assessments and analyzing statistical information is proposed. It is suggested that if there is a priori information about previously occurred events that can be used for risk analysis and fore casting, the fuzzy conclusion should be refined using widely known methods of mathematical statistics, optimization algorithms, for example, gradient descent, simplex method or genetic algorithms. An example of developing a risk assessment system when an enterprise enters into contracts with both the customer and co-executors is given.

risk assessment, fuzzy logic, hybrid neuro-fuzzy inference

1. Zade L. Ponyatie lingvisticheskoi peremennoi i ego primenenie k prinyatiyu priblizhennykh reshenii: per. s angl., pod red. N.N. Moiseeva, S.A. Orlovskogo (The concept of a lingustic variable and its application to approximate reasoning). Moscow: Mir; 1976. 166 p. (in Russ.).

2. Zadeh L. Fuzzy Sets. Information and Control.1965;8(3):338-353. https://doi.org/10.1016/S0019-9958(65)90241-X

3. Grodzenskiy S.Ya., Chesalin A.N. About the usage of fuzzy logic to reliability assessment of automated systems. Nelineinyi mir = Nonlinear world. 2017;5(4):17-23 (in Russ.)

4. Glushenko S.A. An adaptive neuro-fuzzy inference system for assessment of risks to an organization's information security. Business Informatics. 2017;1(39):68-77. http://doi.org/10.17323/1998-0663.2017.1.68.77

5. Chesalin A.N., Grodzenskii S.Ya, Nilov M.Yu. Method of self-assessment of the quality of management decisions. In: Proceedings of the International scientific and technical conference «Fundamental problems of radioengineering and device construction «INTERMATIC–2018»» 2018;(5):1149-1152 (in Russ.).

6. Hajer M., Ketata R., Taieb B., Samir.A. Comparative study of Fuzzy Hierarchical Hybrid approaches for control of Quality Management System. In: International Conference on Industrial Engineering and Systems Management (IESM-2015). 2015. P. 1034-1040. https://doi.org/10.1109/IESM.2015.7380282

7. Ying B., Hanyou W, Longkang W., Kangkang T. Study and analysis on fuzzy quality control for the high- end manufacturing process based on Taguchi quality loss function. J. Comput. Methods Sci. Eng. 2019;19(1):121-136. https://doi.org/10.3233/JCM-180857

8. Grodzenskii S.Ya., Grodzenskii Ya.S., Chesalin A.N. Sredstva i metody upravleniya kachestvom: uchebnoe posobie (Means and methods of quality management: textbook). Moscow: Prospekt; 2019. 128 p. (in Russ.). ISBN: 978-5-392-28446-7

9. Grodzenskii S.Ya. Upravlenie kachestvom: uchebnik. 2-e izd., pererab. i dop. (Quality Management: textbook). Moscow: Prospect; 2089. 320 p. (in Russ.). ISBN: 978-5-392-28172-5

10. Shtovba S.D. Proektirovanie nechetkikh sistem sredstvami MATLAB (Design of fuzzy systems using MATLAB). Moscow: Goryachaya liniya – Telekom; 2007. 288 p. (in Russ.).

11. Goodfellow I., Bengio Y., Courville A. Deep learning. Cambridge: The MIT Press; 2016. 800 p. ISBN: 0262035618.

12. Lokhin V.M., Romanov M.P., Kazachek N.A. The investigation of the periodic oscillations in the control systems with fuzzy controllers. Vestnik MGTU MIREA = Herald of MSTU MIREA. 2015;3-1(8): 138-155 (in Russ.).

13. Myznikova V.A., Ustimenko V.V., Chubar A.V. Fuzzy controllers construction in the SimInTech environment. Kosmicheskie apparaty i tekhnologii = Spacecrafts & Technologies. 2019;3(1):22-27 (in Russ.). https://doi.org/10.26732/2618-7957-2019-1-22-27

14. Kartashov B.A., Kozlov O.S., Shabaev E.A., Shchekaturov A.M. Sreda dinamicheskogo modelirovaniya tekhnicheskikh sistem SimInTech (SimInTech technical systems dynamic modeling environment). Moscow: DMK Press; 2017. 424 p. (in Russ.). ISBN: 978-5-97060-482-3