Uczenie maszynowe jako budulec łańcuchów dostaw odpornych na opóźnienia – podejście zorientowane na interesariuszy
Machine learning as a building block for resilient supply chains against delays—a stakeholder-centric approach

Author(s): Mateusz Wyrembek
Subject(s): Economy, Supranational / Global Economy, Business Economy / Management, Micro-Economics, Management and complex organizations, ICT Information and Communications Technologies, Globalization, Transport / Logistics
Published by: Wydawnictwo Uniwersytetu Ekonomicznego w Poznaniu
Keywords: supply chain resilience;supply chain delays;machine learning;
Summary/Abstract: The author focuses on enhancing the resilience of supply chains in the face of increasing globalisation and the complexity of business processes. The study investigates whether the data collected within the supply chain can be effectively utilised to build systems resilient to disruptions and delays using machine learning methods. Special emphasis is placed on understanding stakeholders’ predictions of delivery delays, which is crucial for maintaining operational continuity and competitiveness. The chapter begins with a literature review on the application of machine learning in supply chain risk management, with a focus on predicting delays. The methodology section presents various machine learning techniques, such as decision trees, bagging, AdaBoost, and random forests. An experiment was conducted on an extensive dataset, using exploratory analysis to identify key features and build classifiers. The study focuses on analysing data from DataCo Global, attempting to predict delays and interpret the results for supply chain stakeholders. The experiment’s results indicate that AdaBoost is the most effective algorithm for this task. This article highlights that machine learning offers promising opportunities in supply chain management but requires continuous development and adaptation to meet the dynamic challenges of this field.

• Details
• Contents