Durlik, Irmina; Miller, Tymoteusz; Kostecka, Ewelina; Pusty, Tomasz; Łobodzińska, Adrianna
(Scientific Journals Maritime University of Szczecin, Zeszyty Naukowe Akademia Morska w Szczecinie,
)
Urban centers, replete with diverse amenities and opportunities, simultaneously grapple with the challenges
brought on by rapid urbanization, notably in the realms of transport and logistics. A pivotal move towards energy-efficient and sustainable systems is essential to mitigate these challenges. In this landscape, machine learning
(ML), and particularly recurrent neural networks (RNNs), emerge as powerful tools for effectively addressing
these urban complexities. This comprehensive review zeroes in on the deployment of RNNs within sustainable
urban transportation and logistics, highlighting their adeptness in processing sequential data, a critical component in various forecasting and optimization tasks. We commence with a foundational understanding of RNNs,
segueing into their successful applications in urban transport and logistics. This review also critically examines
the constraints of current methodologies and potential avenues for enhancement. We scrutinize the application
of RNNs across several areas, encompassing the energy shift in both passenger and freight transport, logistics
management, integration of low- and zero-emission vehicles, and the energy dynamics of transport and logistics. Additionally, the role of RNNs in traffic and infrastructure planning is explored, particularly in forecasting
traffic flow, congestion patterns, and optimizing energy usage. The crux of this review is to amalgamate and
present the existing knowledge on the instrumental role of RNNs in facilitating the transition to energy-efficient
urban transportation and logistics. Our goal is to highlight effective strategies, pinpoint challenges, and map out
avenues for future research in this domain.