Machinery & Energetics

The use of fuzzy cognitive maps for scenario management of biotechnological objects of food production is grounded. It is shown that to use cognitive maps as a tool for aggregating the knowledge of a group of experts, it is necessary to establish the exact value of fuzzy variable relationships between factors, which presents difficulties in creating a cognitive map with a large number of vertices. An algorithm was developed for the creation and practical use of a fuzzy system for generalizing the evaluation of experts in the regular mode. Since one of the main disadvantages of systems based on fuzzy logic is their inability to self-learn and for their adjustment, it is necessary to re-engage experts at full functional stop. The article sets the task of self-adaptation of the fuzzy cognitive map being developed when changing expert estimates or object parameters. To solve this problem, the apparatus of fuzzy neural networks was used. A fuzzy cognitive map was created that operates according to a simplified algorithm of fuzzy inference and allows you to scrutinize the behavior of the system when the values of concepts change. A fuzzy neural network was tested and constructed in MatLAB environment for generalization of expert estimations with corresponding membership functions.

fuzzy cognitive maps, neural network, weights, control scenarios, many concepts