Purpose This paper investigates how a large language model (LLM)-based early-warning system (EWS) for supply chain risk can be systematically aligned with human expert judgement. Building on prior work that introduced the overall system architecture, this study focuses on the detailed prompting logic underlying the automated analysis and on an empirical validation of the resulting AI-based risk assessments. The purpose is to examine how human evaluations can be used to validate and calibrate AI-generated classifications and numerical risk scores, thereby improving their accuracy, reliability and interpretability. In doing so, the paper contributes to a human-centric understanding of AI-supported decision-making in the context of Logistics 5.0. Design/methodology/approach The study draws on two datasets from a real-world validation conducted in January and March 2025, in which AI-generated risk indicators – including impact, likelihood, credibility, sentiment and risk category – were systematically compared with human expert assessments. Deviations between model and expert evaluations were analyzed using established performance metrics such as mean absolute error (MAE), classification accuracy and error distribution patterns. The methodological design follows a human-in-the-loop approach: expert assessments serve as an operational reference standard and are used to iteratively calibrate the prompting logic underlying the automated analysis. Quantitative evaluation is complemented by design-oriented insights derived from the iterative refinement of the system. Findings The findings demonstrate that a structured, iterative human-in-the-loop validation process successfully improves the quality of AI-based risk assessments. In the initial phase, limited contextual grounding led to conservative model behavior, particularly midpoint bias in numerical ratings such as severity, probability and temporal horizon. Prompt refinements systematically integrated supply-chain-relevant anchor examples, partly adapted to company-specific contexts. This enhanced the model's ability to interpret events in operational terms rather than abstract scales. Across the refined prompts, MAE decreased consistently, confirming that context-aware prompt calibration through expert feedback is an effective mechanism for aligning AI outputs with human judgement. Research limitations/implications The validation is based on two empirical evaluation rounds with a limited number of pilot companies, which constrains immediate generalization across industries and regions. Future research may extend the approach to larger datasets, additional sectors and alternative LLM variants. Nonetheless, the study provides a transferable methodological template for validating and calibrating AI-driven risk detection systems in logistics. More broadly, the findings underline the importance of embedding domain knowledge and expert judgement into AI-based classifications rather than relying solely on probabilistic model behavior, thereby supporting human-centric decision-making in line with Logistics 5.0 principles. Practical implications The results support logistics managers in assessing when and how AI-generated alerts can be trusted in operational risk monitoring. By documenting the prompting logic and workflow design, the study enables organizations to implement comparable EWSs using low-code environments. The findings show how structured human feedback reduces noise and misclassification, helping firms focus on truly critical disruptions, avoid unnecessary reactive measures and strengthen resilience and sustainability in supply chain operations. Social implications By aligning AI-generated assessments with human expertise, the proposed system reinforces human oversight rather than replacing it. This supports transparency, trust and responsible AI use in operational decision-making. In line with Logistics 5.0 principles, the approach reduces cognitive overload caused by information noise and enables more balanced, informed and socially responsible responses to supply chain disruptions. Originality/value This paper advances research on AI-based supply chain risk monitoring by providing a transparent and empirically grounded analysis of how LLMs can be calibrated through human expert judgement. Beyond presenting the prompting logic of an operational LLM-based EWS, the study demonstrates how structured expert feedback, context-specific anchor examples and iterative prompt refinement systematically improve model behavior across multiple risk dimensions. In contrast to prior conceptual, simulation-based or purely performance-driven studies, this work shows how and why human-in-the-loop calibration reduces systematic biases rather than merely reporting accuracy gains. By explicitly linking prompting logic, expert validation and operational decision logic, the paper contributes a transferable methodological approach and strengthens the practical and theoretical foundations of human-centric Logistics 5.0 risk management.
Purpose The advent of Industry 5.0 has redefined industrial logistics by prioritizing human-centricity, sustainability and resilience alongside technological advancement. This study aims to identify, structure and prioritize the strategic driving forces of Logistics 5.0 adoption. The resulting framework offers organizations a theoretically grounded and actionable roadmap for navigating this transformative paradigm shift. Design/methodology/approach Driving forces were identified through a comprehensive review of the Logistics 5.0 and Industry 5.0 literature, establishing 18 subcriteria across three main dimensions. Expert evaluations were analyzed using the fuzzy set theory-based Decision-Making Trial and Evaluation Laboratory (DEMATEL) method to map interdependencies and prioritize drivers, with results validated through comparative analyses using the fuzzy set theory-based Analytical Hierarchy Process (AHP) and Full Consistency Method (FUCOM). Findings This study highlights a hierarchical, phased transformation pathway. Human-centric drivers, including workforce safety, lean management and human–machine collaboration, constitute foundational prerequisites. Sustainability mechanisms, notably green logistics and circular economy practices, ensure long-term value creation. Resilience enablers, such as digital technologies and data orientation, build adaptive capacity. Technology adoption depends on human and organizational readiness, while green and circular practices secure enduring operational legitimacy and stakeholder confidence. Originality/value This study establishes a framework for Logistics 5.0 by identifying priority sequences. Human-centric drivers form the essential foundation, sustainability practices direct strategic focus and resilience factors enable operational stability. The analysis confirms that workforce well-being and environmental initiatives must precede the technological expansion. These findings provide managers and policymakers with structured guidance for targeted investment and strategic alignment during the transition to Industry 5.0.
PurposeWhile the importance of logistics and supply chain management in circular supply chains (CSCs) is recognised, current research lacks a granular understanding of the specific capabilities required to translate circular activities into multidimensional value. This study explores the detailed linkage between logistics and supply chain capabilities (LSCCs) and triple bottom line (TBL) value creation across four distinct textile and fashion CSC activities: collection, sorting, reuse and recycling.Design/methodology/approachA systematic literature review (SLR) was conducted on 70 peer-reviewed journal articles published between 2012 and 2025. A thematic analysis was used to inductively identify specific LSCCs within each CSC activity and deductively map their contributions to economic, environmental and social value.FindingsThe LSCCs are synthesised into a two-pillar framework consisting of operational and strategic LSCCs. Findings demonstrate that while LSCCs directly drive economic and environmental value through improved supply chain efficiency and effectiveness, social value is predominantly an indirect outcome fostered through strategic transparency, consumer empowerment and human–machine collaboration in Industry 5.0.Originality/valueThe novelty lies in its granular activity-based approach, revealing that required LSCCs and their value outcomes differ across activities. Unlike a unified view of circular systems, this article systematically links specific LSCCs to holistic TBL value creation, providing distinct pathways to holistic sustainability and clarifies how the inherent challenges between the four CSCs activities can be resolved.
PurposeWarehouse operations are rapidly becoming more robotized to increase performance and reduce costs. A key innovation in the field is human–robot collaboration (HRC). Despite the essential role of human behavior in operational performance, human factors in HRC are still understudied in Operations Management literature. To address this gap, we conducted a unique real-effort experiment in a warehouse especially erected for this study and ground our analysis in Self-Determination Theory, Leader-Member Exchange Theory and Regulatory Focus Theory.Design/methodology/approachThe experiment compares the objective outcomes of collaborative productivity, collaborative accuracy and human pick time between a configuration with the human leading the robot and a configuration with the human following the robot. Additionally, we investigate the behavioral mechanism governing human reactions to the novel collaboration with robots.FindingsWe find that human leading allows for superior collaborative order picking productivity, while human following allows for greater collaborative order picking accuracy. Especially when picking involves traveling between locations, human leading results in shorter pick times. We additionally establish a prevention focus as the factor that allows workers to bridge the productivity gap between the two setups.Research limitations/implicationsA limitation of the lab experiment is its short duration, which necessitates follow-up research in field experiments.Originality/valueThese insights will help warehousing service providers to tailor collaborative robotic solutions based on quantifiable trade-offs between system speed and accuracy.
PurposeThis study questions the need for more visibility to improve supply chain network resilience (SCNR). It investigates how disruptions propagate through real-world supply chain networks and evaluates the effectiveness of different strategies for fortifying key nodes against such disruptions. The aim is to identify practical, data-driven methods that enhance SCNR by prioritising critical nodes for protection using social network analysis (SNA) metrics.Design/methodology/approachAgent-based modelling combined with the susceptible–infected–recovered (SIR) model from epidemiology literature is applied to simulate disruption propagation in ten real-world supply chain networks. Fortification strategies are based on five SNA metrics and evaluated against random node selection. Fortification is implemented by increasing a node's resistance to disruption and accelerating its recovery, an abstract representation of real-world resilience measures such as redundancy, information sharing or collaborative strategies. Each scenario is tested under single-node and multi-node disruption conditions, with 100 repetitions per configuration to ensure robustness.FindingsTargeted node fortification based on SNA metrics significantly outperforms random fortification in reducing performance loss. While page rank yields best resilience benefits on average, simpler metrics like node degree deliver nearly equivalent improvements, demonstrating that effective resilience strategies can be implemented without requiring full network visibility.Originality/valueThis research closes a relevant gap in SCNR literature by validating fortification strategies on realistic, large-scale supply chain networks, moving beyond idealised or synthetic structures. Findings provide actionable, scalable guidance for supply chain practitioners, demonstrating that even basic network metrics enable meaningful resilience improvements in complex supply chains.
PurposeThis study examines the physical internet (PI) concept, characterised as a global, open, interconnected logistics system, as a robust framework for reducing supply chain and logistics risks and significantly improving overall network resilience. PI addresses the systemic inefficiencies and lack of adaptive capability inherent in traditional dedicated logistics structures.Design/methodology/approachThis study identifies a comprehensive set of 26 supply chain and logistics risks across four major categories: supply, demand, operational and external environmental risks. A quantitative decision-making model was employed to assess the capability of the emerging PI paradigm to mitigate these risks. The effectiveness of PI as a risk mitigation solution is assessed by evaluating the four core components of the logistics web, mobility, distribution, realisation, and supply, against the identified risks using the Fuzzy TOPSIS (Fz–Ts) approach.FindingsThe implementation of the PI framework demonstrates significant potential to build resilience, especially against possible supply chain risks. The Fz–Ts analysis, based on expert judgment, quantified the mitigating impact of the PI on major risks. The top three risks prioritised for reduction by PI implementation are logistics outsourcing risks, supplier logistics service risks and risk in custom clearances. The underlying flexibility and greater agility afforded by PI's interconnected logistics services outperform classic models in terms of resilience when facing facility disruptions.Practical implicationsNew PI capabilities are synthesised through the encapsulation component of the logistics network to address supply chain risks. Organisations in the logistics sector can use the results of this study to develop more effective risk management strategies in the context of PI. Organisations will find PI useful for monitoring emerging risks, updating processes and integrating new technologies to stay ahead of potential disruptions to their operations.Originality/valueThis study formalises the risk spectrum of supply chains and their management using PI elements. PI web capabilities, such as realisation, distribution, mobility and supply webs, have been innovatively used to propose risk mitigation and management insights. The new capabilities of PI are synthesised by encapsulating the components of the logistics web to address supply chain resilience.
PurposeThis paper investigates an integration of automation and digital technologies within air cargo logistics, focusing on empirical testing of the O3dyn pallet transport robot prototype at Munich Airport. The presented research identifies challenges and opportunities regarding dynamic airport environments, enhancing understanding of technology implementation and its implications for operational efficiency.Design/methodology/approachThe conducted research employs an empirical methodology involving 10 days of real-world testing. This includes defining test parameters, documenting operational metrics and monitoring interactions toward other robotic systems. Various scenarios were tested to assess the system's effectiveness in navigating complex airport environments, collecting data on travel times, load weights, manual interventions and operational challenges relevant to air cargo handling.FindingsThe findings indicate that while the robot effectively performed transport tasks in a dynamic airport environment, its autonomy was limited, necessitating significant human intervention. Challenges included obstacle detection and navigation, indicating a need for further development in real-time decision-making and integration with logistics processes.Research limitations/implicationsThe focus on a single airport may not fully capture broader challenges, and the short testing duration may overlook various operational scenarios. Future research should involve multiple and diverse environments and longer periods of data capturing for more comprehensive insights into air cargo handling systems.Practical implicationsThis study provides guidance for air cargo logistics stakeholders, highlighting critical investment areas and the need for collaboration among industry partners to overcome automation barriers and challenges.Originality/valueThis paper presents unique empirical findings on air cargo robotics, demonstrating their practical implications and advancing the understanding of automation in dynamic airport environments.
Purpose The ability to replicate global supply chain (SC) structures is important for maintaining and improving the operations of global SCs. This structure, while known to specific manufacturers and key suppliers, is not generally published and may not be totally transparent to all stakeholders. This paper presents a methodology for replicating the structure of global SCs with limited, publicly available data. Design/methodology/approach The proposed SC replication method decomposes the multi-echelon SC into a set of duo-echelon models, each focused on a single commodity, which are reconstituted for a complete representation of the SC for improved computational efficiency. Flows of raw materials, middle products (product components) and end-products for each duo-echelon model are obtained from a developed stacked machine learning method using only publicly available information. The duo-echelon models are integrated into a final multi-echelon, multi-commodity SC representation. Findings The method was applied to a case study of lithium-ion battery production. The findings underscore the model’s capability to capture complex trade relationships and global SC dynamics. Further, the method outperformed numerous other approaches. Originality/value This paper proposes a novel approach to estimating global SC structures, including complexities from their multi-commodity and multi-echelon nature, using publicly available data.
PurposeThis paper investigates the impact of autonomous vehicles (AVs) on the business models of logistics service providers (LSPs) and truck original equipment manufacturers (OEMs).Design/methodology/approachEmploying a multi-case study methodology, this research analyses the business models of 9 LSP firms and 10 OEMs in the EU road transportation sector. Data sources include semi-structured interviews, annual reports and a wide range of industry reports.FindingsThe research reveals a shift in the value network, with OEMs integrating logistics services into their offerings and developing “Network-Operator” models. LSPs, on the other hand, exhibit a reactive approach to AV innovation, focusing on digitalization and internal processes rather than substantial business model transformation.Practical implicationsThe study suggests that LSPs should proactively explore new business models to adapt to the changing landscape. For OEMs, the focus should be on aligning long-term value creation around AVs.Social implicationsThe adoption of AVs represents a shift with social implications, particularly in employment within the transportation sector and the broader societal impacts of enhanced road safety and reduced congestion.Originality/valueThe paper offers novel insights into how AVs are redefining the roles within the road transportation ecosystem.
PurposeRetail logistics is facing higher levels of disruptions, forcing retailers to enhance resilience. However, resilience initiatives such as higher safety stock levels often come at the expense of lower operational efficiency through higher cost levels and/or reduced outcomes.Design/methodology/approachThis study offers a novel approach for addressing this critical resilience-efficiency trade-off based on robust data envelopment analysis (RDEA), combining DEA as a non-parametric tool for measuring efficiency and robust optimization as an established approach for dealing with disruptions. Based on RDEA, this study introduces a metric for quantifying the resilience-efficiency trade-off. Our study empirically validates this novel metric using a representative data set comprising distinct input and output values for stores of a German retail group.FindingsBy empirically proving this metric for a real-world retail logistics case, we show that most stores sacrifice significant portions of efficiency for an increased level of resilience. However, we identify stores that can align resilience and efficiency and are able to provide a refined understanding of how different stores navigate the challenges of a resilience-efficiency trade-off.Originality/valueFor practitioners, this study offers novel insights into assessing the costs of resilience, preventing managers from losing efficiency by prioritizing certain resilience initiatives over others (e.g. logistics activities over personnel resources). These insights enable a more balanced perspective in strategic resilience decision-making and logistics management.
PurposeA paradigm for Puzzle-Based Storage Systems (PBS) is that all grid cells have the capacity to move loads horizontally and vertically. This research challenges this paradigm by considering unidimensional PBS (UPBS), where some grid cells are limited to either horizontal or vertical movements. UPBS simultaneously reduce the investment cost and operational complexity of PBS, at the expense of a potentially lower system throughput.Design/methodology/approachA linear program (LP) formulation is presented to determine the optimal retrieval path for a load using a single escort (i.e. an open cell in the grid that allows load movement) in a UPBS. The LP is solved recursively to understand the tradeoffs of unidimensional designs for a 4 × 4 system. Formulations to solve the single-load multi-escort problem are also proposed. Lastly, an existing decentralized control PBS algorithm was used to develop managerial insights for designing UPBS, considering the simultaneous retrieval of multiple loads using multiple escorts and input/output points.FindingsIt is concluded that UPBS with three unidimensional cells would reduce the PBS cost by 1.8 times the capital cost of a bidimensional cell, at the expense of an 8.34% average reduction in throughput. It is argued that multi-escort UPBS designs should be similar to the ones empirically produced for single-escort UPBS. Lastly, it is concluded that UPBS may provide a favorable capital investment cost-to-throughput tradeoff.Originality/valueThis study is a pioneer in considering UPBS. The empirical evidence of a favorable system cost-to-throughput ratio of UPBS improves the business case of implementing PBS in practice.
PurposeIn early 2020, the coronavirus disease 2019 (COVID-19) pandemic reached global proportions within only a few weeks. A key strategy to contain and control pandemics is to isolate infected people, which requires making test results available quickly, in order to be effective. The contagious disease testing problem (CDTP) arises precisely in this context of testing potential cases (of infection). In this paper, we address the stochastic and dynamic version of the CDTP, where only some suspected cases are known in advance, while others arrive randomly throughout the course of the day. For each newly arriving suspected case, it must be decided whether to accept or reject it. The specimens of accepted suspected cases must be collected on the same day; either by assigning the case to a time slot in a test-center or by visiting the patient with a mobile test-team. Rejected suspected cases must be serviced on the next day. The task in this problem is to decide how many mobile test-teams to use, how many test-centers to open and where, which suspected cases to visit with a mobile test-team and which to assign to a test-center, and designing the vehicle routes for the mobile test-teams. The objective is to identify a dynamic assignment-and-routing policy that minimizes the number of open test-centers and used vehicles, services all early-known suspected cases, and maximizes the expected number of serviced late-known suspected cases.Design/methodology/approachWe introduce this new problem, which we call the stochastic dynamic contagious disease testing problem (SDCDTP), and formulate it as a Markov decision process. We propose a solution method based on value function approximation for solving the SDCDTP, and illustrate its performance with an extensive computational study.FindingsWe solve and analyze real-world-based problem instances from which we draw insights regarding the solution quality produced by our approach.Originality/valueWe find that a (significant) increase in the number of covered late-known suspected cases can be achieved by not minimizing the number of used vehicles and test-centers, albeit for the price of equally significant increases in the amount of used infrastructure.
PurposeThe paper aims to explore the use of an augmented reality assistance system to enhance safety in forklift operations by addressing visibility challenges caused by structural and load-induced obstructions. A user study validates the potential of see-through as a driver assistance system on forklifts, and highlights challenges in accuracy, latency and image quality.Design/methodology/approachThe paper opted for a comprehensive user study with 18 participants. The tests were conducted while the forklift was not driving. Participants were allowed to elevate and incline the mast to evaluate the system’s performance. All participants answered 11 questions focusing on performance, usability and optimization. The questionnaire used an 11-point Likert scale (0: strongly disagree, 10: strongly agree) to quantify the personal opinions of the users. Depending on the hypotheses, one-sample or two-sample t-tests were carried out to evaluate the questionnaires.FindingsThe paper provides insight into the evaluation of an augmented reality-based driver assistance system for forklifts. The study identified latency and image quality as critical areas for improvement, essential for enhancing user acceptance and operational reliability.Research limitations/implicationsBecause of the restriction of the participants to a university environment and the fact that the user study was carried out while the vehicle was not driving, the research results may lack industrial transferability. For this reason, researchers should carry out further investigations in companies, for which considerable safety precautions are necessary.Originality/valueThis paper fulfills an identified need to study how the visibility problem on forklifts can be solved and thus increase safety in intralogistics.
Purpose The dynamic nature of production environments with uncertainty about future tasks places a high demand on the flexibility for assigning tasks to a fleet of Automated Guided Vehicles. The goal of our research is to reduce the number of required vehicles compared to the traditional approach while achieving the same throughput for economic and sustainability reasons by enabling a more intelligent battery management. Design/methodology/approach Our proposed approach considers charging as an option during robot task assignment. We extend a previously presented sequential single-item auction algorithm for online scheduling using utility functions to compare the options for performing tasks or charging. This enables a more intelligent decision, depending on the current system state, focusing on improved availability in phases with high demand and efficiency. A material flow simulation highlights potential issues when disregarding battery management and validates the proposed methods. Findings Despite the advancements in modern battery technologies, battery management remains an important aspect for continuous operation without breaks for charging in between shifts. Additionally, frequently driving to charging stations for opportunity charging can generate significant additional movement and power consumption. It is demonstrated that our approach improves system performance while reducing the total number of charging operations, which allows a reduction in required vehicle count. Originality/value Our approach represents a deeper integration of battery management aspects into the robot-task assignment problem. The proposed online scheduler is very flexible, as any number of system states can be considered in the form of additional utilities, enabling potential for further optimization.
Purpose Sustainable Aviation Fuel (SAF) is crucial for aviation decarbonization, but its current pre-blending process at refineries presents challenges, including fixed blending ratios, higher transportation costs and long lead times. This study explores the potential of an innovative technology that enables on-site SAF blending at airports. By postponing blending to the point of use, this approach offers customization opportunities. However, the precise benefits and trade-offs of this concept remain unclear. The research aims to assess the impact of on-site blending on fuel price, lead time, carbon emissions and supply chain costs. Design/methodology/approach This empirical study evaluates the effects of SAF postponement using case analyses of Singapore-Seletar and Maastricht airports. The analysis incorporates cost modeling, lead time assessment and carbon impact calculations to quantify the implications of shifting blending downstream to airport sites. Data sources include industry reports, airport-specific logistics information and SAF supply chain parameters. A comparative analysis is conducted to determine optimal airport conditions for SAF postponement, highlighting key enablers and barriers to implementation. Findings The results indicate that on-site SAF blending can create competitive advantages by reducing supply chain costs and lowering carbon emissions. The benefits are contingent on airport-specific factors, such as Hydroprocessed Esters and Fatty Acids availability, logistics infrastructure and regulatory conditions. The findings suggest that certain airports, particularly those with strategic locations and favorable cost structures, are better suited for adopting SAF postponement. By shifting production downstream, airports can achieve greater flexibility in SAF blending ratios while minimizing logistical inefficiencies. Originality/value To the best of the authors’ knowledge, this study is among the first to empirically examine the feasibility of postponing SAF blending to the airport level. While existing literature focuses on SAF production and distribution, the concept of downstream blending has not been systematically analyzed. The research provides new insights into how mass customization principles can be applied to SAF supply chains, potentially reshaping fuel logistics in the aviation industry. By identifying critical factors for successful implementation, this study contributes to both academic discussions and practical decision-making in sustainable aviation fuel management.
