Application of Artificial Intelligence Methods for Inventory Management Efficiency Improvement

The key function in logistics is the inventory management function. The importance of improving the efficiency of decisions in the field of inventory management is beyond doubt: the stock is a material flow at rest, i.e. Is a frozen money, with inefficient decisions, the stock rises, which threatens to overflow the warehouse space, the deterioration of financial performance and losses in the form of write-offs due to the expiration date. In accordance with the system approach, when making decisions it is necessary to consider how these decisions affect the entire system as a whole, therefore, the effectiveness of decisions in inventory management is important from the supply chain viewpoint. To make optimal decisions specialists in companies need to analyze large amounts of different data, inventory management is no exception: in addition to the amount of free space in the warehouse, the time of delivery, the inventory manager should take into account additional factors related to the consumption of stock or the duration of logistics cycles, In words, an expert must have forecasting skills. The modern level of computing capabilities of computers and the level of development of the theory of artificial intelligence makes it possible to create a special class of programs called "intelligent agents" that make decisions in accordance with specified functions. Agents are used in many areas of knowledge and business, but no attempts have been made to use intelligent agents to solve logistics problems. The paper presents the mathematical function of the intelligent inventory manager developed in the programming language Python 3.5, consisting of a stock management model in a 2-level integrated supply chain and a combined predictive model based on an artificial neural network and 6 simple prediction methods. Evaluation of the agent's actions showed the effectiveness of the decisions taken in comparison with the Aksater model.