Increasing the efficiency of warehouse analysis using artificial intelligence

Abstract

Logistics in companies is a necessary process that has high costs with mostly no added value. Lowering this cost is vitally important for companies to stay competitive. Nowadays, storage systems are a critical part of any company’s logistic system, and many of them try to reach an optimum level where they can operate with little freedom of movement of goods declared by the changing market. There are several manual and automated methods to achieve this. However, we hear quite little about the use of artificial intelligence in the field. This study focuses on the implementation of AI technology into warehousing, especially in categorizing goods. After an overview of the recent literature on AI technologies and their application in the field of logistics, the introduction of an AI application follows. The main goal of the application is to categorize each good stored in a warehouse into ABC-XYZ groups, which determines the place of the good in the warehouse and the ordering frequency with the quantity. After acquiring and cleaning the training data from a real company, the determination and selection of the least input parameters is an important and challenging task, which is demonstrated. The effectiveness of the supervised learning can be seen as an ANN (artificial neural network) can output, with the aid of a non-conventional metaheuristic approach - the black hole algorithm - as the learning agent is demonstrated by an example, which also shows the result of an ABC-XYZ categorization run on a dataset from a multinational company.