به كارگيري يادگيري ماشين در بهينه‌سازي مصرف مواد اوليه در برنامه‌ريزي توليد محصولات برودتي

In todayʹs industrial world, optimizing raw material consumption is one of the main challenges in production planning. The company that produces refrigeration products, Haier asa, in Entekhab Industrial Group (Snowa), which is one of the leaders in the home appliance industry, is facing problems in optimizing an‎d managing raw material consumption. This thesis examines an‎d analyzes the application of reinforcement learning techniques as a new solution for optimizing raw material consumption in the production planning of refrigeration products. In this research, we analyze data using the Python programming language. The main goal of this research is to develop an intelligent model that can provide accurate planning for the production an‎d consumption of raw materials that remain as an inventory of production deviations by analyzing the data available in the system. It is also important to analyze the compliance of the obtained production plan based on the real potential of resources (raw material inventory) with actual production. Using the proximal policy optimization algorithm in machine learning, it is expected that this model can significantly increase the effectiveness of production processes an‎d introduce a data-based approach an‎d scientific analysis instead of traditional an‎d empirical methods. The results of this research can help improve the performance of Entekhab Industrial Group (Snowa) in the refrigeration products of Hair asa Company an‎d increase its competitiveness in the market. In accordance with the model assumptions, the average monthly profit reached more than 55 billion Tomans, 100% efficiency in matching production with sales, an‎d the customer service level was also achieved at about 80%, which indicates the modelʹs power in making correct decisions an‎d the modelʹs rational performance under resource constraints.

5