• Some road freight transportation decisions remain uninvestigated. • Paucity of application to real-world road freight transportation decisions. • Operationalisation of a framework to investigate data-driven decisions. • A possible theoretical lens is proposed to further expand the research on the topic. Road freight transportation represents the dominant mode for inland freight transportation and is a critical component of logistics systems. Decisions in this domain are inherently complex due to the scale of operations, the numerous actors involved, and the dynamic operating environment. Machine learning (ML) represents a valid tool to address this complexity. However, literature on the topic is largely technical and focused on specific applications, lacking broader managerial insights that researchers and practitioners can leverage when applying ML to road freight transportation management (RFTM) decisions. To address this gap, this paper presents a systematic literature review investigating the characteristics of decision making when ML is applied to RFTM, encompassing both the managerial and technical aspects. The findings indicate that, while numerous studies focus on vehicle routing problems, other decisions within RFTM remain largely underexplored. Furthermore, only a limited number of studies develop and validate machine learning (ML) solutions using real-world data, thereby constraining the assessment of their practical applicability and impact. In addition, drawing from the analysis of the review results, this study proposes a conceptual framing of ML applications in RFTM, depicting ML as a potential mitigator of the limitations associated with bounded rationality in RFTM decision-making. This paper lays the foundation for future research on the application of ML for RFTM. It also proposes a replicable approach for analysing data-driven decisions in other contexts. Moreover, it offers practical guidance to decision makers by highlighting the elements that ML can act upon, thus supporting its adoption in practice.
Mascheroni et al. (Wed,) studied this question.