DataRel - Big Data for Resilient Logistics

DATAREL aims at advancing the extant logistic knowledge with novel Internet of Things (IoT) and Big Data solutions for detecting emergence, which is used to improve the quality control and multimodal planning in terms of resilience, real time, efficiency, and dynamics. In particular, DATAREL addresses a system comprised of two main intelligent components: the frontend Logistics Internet of Things (LIOT) and the backend Artificial Logistics Intelligence (ALI). In the frontend, LIOT investigates a radically new approach to logistic processes involving cyber-physical networks of logistic entities (e.g., goods, vehicles, infrastructure) that embed smart sensors and business logic, forming thus Smart Returnable Transport Items (SRTIs). SRTIs are capable of collecting data from extant IoT devices and social networks. LIOT can detect unexpected behaviour and do adaptive sampling of data generated by collaborating SRTIs in the region of interests. Such capabilities enable the extension of logistic processes with the real-time monitoring and dynamic planning capabilities at the point of action. In the backend, ALI is responsible for logistics planning given subsequent masses of data provided by LIOT. ALI also has learning capabilities to self-evolve with unprecedented emergence from the dynamic environment through interacting with human planners. We argue that such an intelligent system significantly improves the real time and resilience of planning decisions under uncertainty and time pressure. Compared to existing prescriptive approaches, our online predictive/descriptive approach is much better suited to deal with current challenges of IoT and Big Data such as volume, trustworthiness, heterogeneity, velocity, and dynamics. The implementation of this concept relies on a common cross layer service paradigm, going from enterprise applications down to the logic executed on SRTIs (e.g., sensors and actuators). We choose two business scenarios as frame of reference for the DATAREL research, each posing significant challenges: the real-time monitoring of shipments, and the real-time multimodal planning. To address practitioners’ needs resulting from these scenarios we refine DATAREL’s research goal into two themes as follows: 1) Real-time perishable shipments monitoring with the advances of IoT: A focal point of this theme is aimed at an optimal solution for the problem of lost, depreciated, and damaged perishable goods during transportation and storage. Hence we propose LIOT to enhance the IoT sensing capability, and device-crossing platform for real-time monitoring, detection, analysis, and prediction of logistic emergent behaviours. 2) Real-time resilient multimodal planning with the advances of Big Data: The main goal is to improve the resilience and performance of planning decisions in terms of real time, sustainability, efficiency, queue delay, and quality. To this end, we propose a multi-agent ALI that uses the detected sources of emergence to improve tactical, and operational decision-making through agent-based modelling and simulation. ALI also can learn from the creativity and experience of human planners using advanced Machine Learning techniques. Through these themes, this project brings together a critical mass of world-class research teams to aim at a smart logistics system that can handle the dynamics, uncertainty, and time pressure. The partnership will build integrated logistic networks and develop novel behaviour recognition and data mining algorithms.