Guiding practitioners of road freight transport to implement machine learning for operational planning tasks

Road freight transport is the most used mode, and its importance is expected to increase. Access to any location is the main reason for its proliferated use. However, the industry faces myriad challenges due to acting in a complex, dynamic environment. Effects, such as rising customer expectations or the need for sustainability, increase the pressure for decision-making. Operational planning is particularly affected due to numerous decisions and short lead times. Furthermore, decisions can become unfavorable within minutes due to external developments, such as traffic jams. Technological developments have led to storing more data. Machine learning, is regarded as powerful to generate valuable insights for road freight transport. However, despite some successful lighthouse projects, practitioners struggle to identify use cases. Even though information on operational planning is available, a lack of machine learning knowledge hinders implementation. To support road freight transport practitioners in reaping the envisaged benefits, a machine learning implementation guideline is proposed. Design science was used to develop the guideline, established on a multi-vocal literature review. Existing frameworks and their drawbacks were identified to build a guideline addressing all phases relevant to machine learning projects. The guideline was evaluated with a use case from a road freight transport company.