Our ambition is to promote the availability of high quality fresh food with low environmental footprint, using a data-driven approach.
The complexity of the fresh food supply chain is increasing. It is a major challenge to ensure optimal product quality and availability in a sustainable and cost efficient manner. Current approaches to supply chain control and optimisation are typically based on practical experience and intuition, even in modern control systems. The limits of this approach are in sight. With Big Data technology rapidly evolving, now is the right moment to make a leap forward. It is time to begin exploiting the data that is already being generated in a diversity of sources, from high-tech sensors to written notes. Combined with expert knowledge, this will facilitate supply chains to follow entirely new operational strategies.
This project aims to implement this Big Data vision in the Dutch fresh product supply chain, for now focussing on the optimisation of the conditions in which products are stored and transported. Dutch companies in this sector are leading world-wide and already generating massive amounts of data on pre- and postharvest conditions and product quality. However, the wealth of information hidden in this data is not yet used to create advanced, evidence-based control mechanisms. Moreover, gaps still exist in the data collection process.
The Cool Data Hub
In this project we want to prove that Big Data technologies (in particular machine learning combined with linked data standards and knowledge modelling) will allow storage and transport providers to operate significantly more effective and efficient than they do now. Ultimately, the partners in this project aim to establish an open platform in which the developed data, models and algorithms can be continuously shared and updated: the Cool Data Hub.
This project will lead to improved and homogeneous product quality and a more sustainable process in the fresh product supply chain. The project will create new business opportunities throughout the chain and in IT. Scientifically, our challenge is to integrate self-learning methods such as Bayesian Belief Networks with Linked Data, i.e., semantically enriched data. If this can be done, a continuous self-learning cycle involving data, models and applications can be realized. Moreover, the generated data will allow agrifood scientists to create new hypotheses for further experimental research.