METHODOLOGY FOR DETERMINING THE TIMING OF THE MANUFACTURE OF NEW ORDERS IN SINGLE PRODUCTION AND REDUCING THE RISKS OF PRODUCTION ACTIVITIES

The article is devoted to the development and testing of the methodology for assigning the priority of technological operations for a number of orders and assessing the risks of setting the price and deadlines for new orders, taking into account it’s complexity and priority. It is noted the main problems of high-tech single production and the risks they create. To solve the identified problems on the basis of the analysis, a complex method was chosen, consisting of a combination of the use of simulation modeling and neural network modeling. The neural network model is based on the statistics of the production times of various parts in this area. The type of neural network model selected by the Fitting app is trained by the network using the Levenberg-Marquardt error back propagation algorithm. The simulation model of the production site is built in the Tecnomatix Plant Simulation program. As a result, thanks to the developed methodology, it became possible to obtain information about a new order before it was directly introduced into production, to diversify risks before they caused damage, as well as to improve reputation. In conclusion, as an example, the addition of a new part to other parts already being produced is given, and the time of its production is calculated.