Hybrid model of an expert system for assessing the stability of a production system

• Main Authors: Anna Burduk, Katarzyna Grzybowska, Gábor Kovács
• Format: Article
• Language: English
• Published: Poznań School of Logistics 2018-12-01
• Series: LogForum
• Subjects: production system; risk assessment; artificial neural networks; fuzzy logic; stability; variability
• Online Access: http://www.logforum.net/vol14/issue4/no6/14_4_6_18.pdf

Background: The article presents the concept of control of the production system, which allows to maintain its stability, and thus to implement the established production plans. For this purpose, combinations of simulation models and artificial neural network (ANN) models of the production system have been suggested. The combination of both types of models was possible thanks to the development of a hybrid model of the expert system to assess the possibility of implementing the production plan (objective) depending on the risk size and the level of stability of the production system analysed. The analysed problem - the possibility of implementing production plans depending on the risk size and the level of stability of the production system - is difficult to mathematical modelling. However, based on the data analysis from the simulation model and the ANN model, we can obtain information on the dependences of the corresponding input and output values. Methods: Based on the presented method of managing the production process using computer models, the possibilities of using simulation models and ANN models in assessing the stability and risk of production systems have been analysed. The analysis and comparison of both types of models have been performed due to the construction and the type of input and output data. Results: The direct combination of simulation models and ANN models is not allowed by their different structure, specificity and other types of input and output data. Therefore, the concept of combination of both types of models presented in the article is conducted via a database of expertise and fuzzy inference. Conclusions: For the purpose of controlling the production system, it was suggested to build a hybrid model of an expert system to assess the possibility of achieving the objective depending on the risk size and the level of stability of the production system.