Development of forecasting and management models for automated production systems

Abstract

In modern conditions of dynamically developing production and increasing competition in the market, an urgent task is to develop effective forecasting and management models for automated production systems (APS). This research is aimed at creating an integrated approach to modeling and optimizing the functioning of APS in order to increase the efficiency of production processes, reduce costs and ensure high quality of products. To achieve these goals, methods of mathematical modeling, control theory, optimization and data mining were applied. In particular, stochastic models for forecasting product demand have been developed, taking into account seasonal fluctuations and market trends. Such models allow predicting sales volumes for a period from 1 to 6 months with up to 95% accuracy. Adaptive scheduling and dispatching algorithms based on fuzzy logic and genetic algorithms have been proposed for managing production processes. Using these approaches made it possible to reduce the time of equipment changeover by 20-25% and reduce the volume of work in progress by 15%. The conducted experimental studies on the example of a real machine-building enterprise have confirmed the effectiveness of the developed models and algorithms. The implementation of the proposed solutions allowed to increase the productivity of the APS by 12%, reduce the cost of raw materials by 8% and improve the quality of products, reducing the percentage of defects from 1.5% to 0.8%. The results obtained are of high practical importance and can be used to improve the competitiveness and efficiency of enterprises in various industries. Further research will be aimed at developing the proposed approaches and adapting them to solve new problems in the context of digitalization of production and the transition to the concept of «Industry 4.0».