Author: Regina Enrich Sard, Giulia Zarpellon and Laia Garriga, EURECAT
The urban distribution of goods encompasses the entire flow of merchandise that occurs within a city. This includes not only the final step of the supply-chain (i.e., its last mile) but all the operations associated with delivery, collection, transfer, loading and unloading, placement at points of sale, and return of reverse logistics.
The whole process is a key axis for cities, given that it allows the distribution of basic services to the urban population while supporting the development of local economies. At the same time, however, it generates important externalities – like environmental and acoustic pollution and the occupation of public space. With e-commerce expansion and consumers’ demands increasing (for example with respect to delivery times), the sector and the territory where it operates are put under great pressure, and generated impacts are felt more and more. For example, in 2018 Barcelona made 23 million online purchases, with 86% of deliveries made to homes or offices. At the level of Catalonia, in the period from 2018 to 2020, there was an increase in business-to-consumer trade of 44.6% and an increase of 15% in home deliveries (www.barcelona.cat). The management of the urban distribution of goods represents a challenge for any municipality that, without affecting well-being or reducing services, wants to improve in terms of sustainability, efficiency, and safety.
Today, urban distribution relies on highly digitised processes, which generate significant amounts of data. The availability of these data outputs creates the perfect ground for the application of technologies like Artificial Intelligence (AI).
In a nutshell, AI provides techniques and methodologies designed to allow computer systems to simulate human intelligence processes, such as recommendation and decision-making for optimal scenario management. Within AI, the field of machine learning is focused on developing strategies to generate knowledge from previous experiences. In a setting with high availability of historical information like logistic operations, machine learning algorithms can enable us to learn from past experiences by exposing patterns in the data and potentially correlating it with past events. Data can also be fed to predictive systems to forecast what the most probable future will be at different time scales, while the assessment of what-if scenarios can support decision-makers with data-driven and more informed alternatives in the face of unforeseen situations.
A core potential of generating useful knowledge with AI-approaches lies in the joint use of heterogeneous data sources – with administrations, companies, and society exchanging and acting upon combined information in almost real time. For example, data from different public and private sources can serve as a basis for generating urban logistics behaviour models and can be enriched with other data of interest like traffic information, weather forecasts, or current regulations. In this context, the ability to share and collaboratively utilise diverse types of information in secure manners is crucial.
The GREEN-LOG project encompasses all these themes and aims to create participatory solutions towards more sustainable and data-driven urban distribution systems. It is now essential for these technologies to be available to all players in the value chain, regardless of their size or capacity. Only in this way we will be more efficient and resilient in the highly dynamic and changing environment we live in.
Picture source: NaruFoto/Shutterstock.com/