Maria Boluda-Prieto

Integrated berth allocation, quay crane assignment and scheduling constitute an established approach to coordinating seaside container-terminal operations. This study extends this research stream by incorporating workforce planning and economic, environmental and social sustainability objectives, constraints and decision variable by means a multi-objective mixed-integer nonlinear programming model is proposed The formulation is linearised and converted into a normalised weighted-sum mixed-integer linear programming model. A novel three-phase matheuristic combining multi-start constructive generation, reference-solution selection and restricted MILP reoptimisation is also proposed to accelerate solution generation. Computational experiments using literature-based benchmark instances validate the objective structure, analyse sustainability trade-offs, examine the impact of quay-crane capacity and evaluate scalability under increasing vessel demand. The results show that excluding one sustainability dimension can substantially deteriorate the remaining performance indicators, whereas assigning positive weights to all objectives avoids the most extreme operational outcomes. Increasing quay-crane capacity produces marked improvements up to 12 cranes, after which marginal benefits diminish. For computationally demanding instances, the matheuristic obtains feasible solutions in approximately five minutes, compared with the two-hour limit of the complete MILP. For instances with 6–18 vessels, its mean solutions remain within 2–7% of the MILP incumbents. It obtains better objective values than the MILP incumbents in several larger scenarios and returns feasible solutions for 26–30 vessels, for which the MILP finds none within the time limit. These findings demonstrate the computational value of the matheuristic when unrestricted MILP optimisation becomes impractical, or a rapid solution is required thereby supporting operational resilience under high-demand and time-constrained conditions.

Industry 5.0 demands that engineering graduates develop competencies that integrate technical knowledge with human-centricity, sustainability and resilience. However, engineering education has not always advanced at the same pace, and effective educational designs that operationalise these principles within applied learning contexts remain scarce. This paper presents Innovation Day, a four-hour challenge-based learning experience that operationalises Industry 5.0 principles through an artificial intelligence (AI)-supported cyber–physical logistics challenge in management engineering master’s programmemes. The activity engaged 47 students organised into 14 teams across two master’s degrees at the Universitat Politècnica de València. Students progressively developed and integrated data structuring, QR-based identification, sequencing and storage assignment logic, computer vision and collaborative robotics within a single supervised automation workflow. All 14 teams completed the first three phases, while completion rates reached 71% in Phase 4 and 64% in Phase 5, reflecting the increasing demands of system integration. Results also show that generative AI acted as a pedagogical tool, enabling students with limited programming experience to develop functional Python applications while maintaining human validation and accountability. The experience demonstrates that complex Industry 5.0 scenarios can be translated into manageable, transferable educational activities through phased challenge design, functional validation gates and structured AI-supported learning.

The surge in containerised trade has intensified the need for efficient resource allocation in port operations, particularly in berth allocation (BAP), quay crane assignment (QCAP) and quay crane scheduling (QCSP) problems. While mathematical programming and metaheuristic approaches have traditionally been used to solve these problems, their scalability and adaptability remain limited. Recent advances in Machine Learning (ML) offer new optimisation possibilities. This paper conducts a systematic literature review of approaches that apply machine learning techniques to seaside port operations (BAP, QCAP, QCSP, and their integrated variants). Operations such as quayside transport planning, landside or yard activities are outside the scope of this review. The review was conducted following PRISMA guidelines and was based on publications indexed in Scopus and Web of Science. The selected works are analysed and classified across nine dimensions. The results highlight a predominant focus on economic optimisation, with BAP being the most frequently addressed problem. Earlier studies broadly applied regression techniques and predictive approaches. At the same time, more recent research has shifted towards deep reinforcement learning and prescriptive solutions, opening new opportunities for sustainable and adaptive port operations. Finally, it has identified key trends and opportunities in ML applications for port operations.

The rapid integration of generative artificial intelligence (AI) into higher education raises questions about its pedagogical value and its role in sustainability-oriented learning. This study investigates the use of AI in a Challenge-Based Learning (CBL) engineering hackathon (INDUSHack) focused on the Sustainable Development Goals (SDGs), examining patterns of AI adoption, perceived benefits and risks, and its contribution to SDG integration in project design. A descriptive exploratory design was employed, collecting questionnaire data from 60 undergraduate participants and complementing these with expert committee evaluations of project outcomes. Results indicate widespread AI use, primarily for idea generation, information synthesis, and project structuring. Participants reported perceived improvements in efficiency, creativity, and the clarity of SDG alignment; however, difficulties persisted in translating sustainability concepts into measurable, technically grounded criteria. The findings suggest that while AI can support learning processes and sustainability framing in CBL environments, its effective use requires structured guidance and critical AI literacy to ensure meaningful integration of SDGs.

The continuous growth in the volume of goods transported by sea has led to a significant increase in the number of vessels and containers arriving at ports, requiring more efficient management of vessel loading and unloading operations. These operations involve three key problems: berth allocation, quay crane assignment, and quay crane scheduling. In response to these challenges, numerous mathematical programming models have been proposed, and various literature reviews analysing these models have emerged in recent years. However, the abundance of such reviews highlights the need for a systematic organisation of available information to consolidate the state of the art in this field. To this end, this paper conducts a systematic literature review, following PRISMA guidelines, based on publications retrieved from Scopus and Web of Science. This study focuses specifically on existing state-of-the-art studies that analyse mathematical models for berth allocation, quay crane assignment, and/or quay crane scheduling. The objectives of this work are threefold: (i) to assess the scope of previous literature reviews, including their purpose, the problems addressed, and their temporal coverage; (ii) to identify and group the dimensions they use to categorize the problems and the models designed to address them; and (iii) to compile the future research directions proposed in these reviews and highlight those that remain relevant. The findings provide a comprehensive overview of trends, highlight research gaps, and propose dimensional groups to consider when developing more holistic decision support frameworks for port operations.

The increasing complexity of IT infrastructures, driven by the rise of IoT devices and microservices architectures, has highlighted the need for robust and scalable solutions to ensure network resilience. This study aims to guarantee system continuity in network failures by proposing a self-healing architecture that integrates monitoring, visualisation and orchestration tools for Kubernetes management. An architecture is proposed that integrates different microservices that collect real-time metrics, visualise data, generate alerts and apply corrective actions. Unlike traditional approaches limited to passive monitoring, an automated, rule-based mitigation flow is introduced, enabling detection and visualisation of failures and the implementation of corrective actions in real-time. A comparative analysis is also included with traditional approaches, highlighting key advantages such as reduced response times, elimination of human errors, better resource management and scalability in dynamic environments. Finally, limitations related to hardware capacity and the management of simultaneous alerts in multi-cluster infrastructures are identified.