Analysis of practices in applying artificial intelligence and machine learning to improve urban route planning results

DOI: 10.31673/2412-9070.2023.049835

Abstract

Against the backdrop of the COVID-19 pandemic since 2019, characterized by the implementation of stringent quarantine restrictions in public cities and a general fear of virus transmission, there has been a significant increase in global demand for e-commerce services [8–11], particularly doorstep delivery of goods. The practice of delivering goods directly from stores to doorsteps, even within city neighborhoods, has gained widespread adoption [9; 10], especially in developed countries. Consequently, the market for courier and postal services has substantially grown [12; 13].
The active use of courier services and the relaxation of quarantine measures, leading to a return to pre-pandemic levels of public and personal transport usage, have overtime increased the strain on urban road infrastructure. While the purchasing power and attitudes of consumers toward large expenses have decreased due to rising global oil prices and, accordingly, increased fuel costs [14], the demand for courier and postal services as a whole has remained steady. This is attributed to the growing trend of demand for the delivery of essential food and goods amid restrictions, as was seen during the pandemic [9]. Additionally, competition in this market is continually on the rise [15]. These factors pose a series of challenges for service providers, one of which is improving the methods and algorithms for calculating delivery routes within cities, as many of these providers also cover last-mile delivery. This is primarily necessary to minimize ancillary costs (fuel, vehicle maintenance, etc.) and, consequently, to offer a better proposition to users.
In most cases, the calculation of delivery routes resembles the classic "Traveling Salesman Problem." Most existing algorithms for route calculation rely on static data. Today, such calculations for city deliveries can quickly become outdated due to various factors such as traffic jams caused by various circumstances, adverse weather conditions, and more. The main challenge with classic algorithms is the need for real-time route adjustments, which require constant recalculations. Although classic algorithms efficiently compute optimal routes, they are quite slow when it comes to adjusting routes in real-time.
One solution to address the issue of efficient route calculation is the application of Artificial Intelligence (AI) and Machine Learning (ML) techniques and incorporating ML elements into the adjustment of calculation results through conventional algorithms or fully relying on AI for efficient route planning.
This research is conducted in the context of C2C (Customer to Customer) delivery of small-sized cargo. The possibility of improving the calculation of efficient routes between pick-up and drop-off points on a single route is explored. This scheme involves parcel collection and delivery as an ancillary process for the courier, such as in scenarios where an individual wishes and is able to deliver some parcels along their route to compensate for their personal transportation expenses. Therefore, this study examines the possibilities and practices of enhancing the calculation of such routes using AI and ML based on previous related research.

Keywords: artificial intelligence; machine learning; last-mile delivery; logistics; real-time calculations; C2C.

References
1. Kervola H., Kallionpää E., Liimatainen H. Delivering Goods Using a Baby Pram: The Sustainability of Last-Mile Logistics Business Models. URL: h t t p s : / / d o i . o r g / 1 0 . 3 3 9 0 / s u 1 4 2 1 1 4 0 3 1 (22.10.2022 р.)
2. Collaborative Mechanism for Pickup and Delivery Problems with Heterogeneous Vehicles under Time Windows / Y. Wang, Y. Yuan, X. Guan [et al.]. URL: https://doi.org/10.3390/su11123492 (25.06.2019 р.)
3. Recent advances in hybrid priority-based genetic algorithms for logistics and SCM network design / M. Gen, L. Lin, Y. Yun, H. Inoue. URL: https://doi.org/10.1016/j.cie.2018.08.025 (2018 р.)
4. Singh A., Wiktorsson M., Hauge J. B. Trends In Machine Learning To Solve Problems In Logistics. URL: https://doi.org/10.1016/j.procir.2021.10.010 (2021 р.)
5. Bosse S. Self-adaptive Traffic and Logistics Flow Control using Learning Agents and Ubiquitous Sensors. URL: https://doi.org/10.1016/j.promfg.2020.11.013 (2020 р.)
6. Knoll D., Prüglmeier М., Reinhart G. Predicting Future Inbound Logistics Processes Using Machine Learning. URL: https://doi.org/10.1016/j.procir.2016.07.078 (2016 р.)
7. A machine learning optimization approach for last-mile delivery and third-party logistics / M. E. Bruni, E. Fadda, S. Fedorov, G. Perboli // URL: https://doi.org/10.1016/j.cor.2023.106262 (2023 р.)
8. Covid-19 Pandemic as Sustainability Determinant of e-Commerce in the Creation of Information Society / W. Chmielarz, M. Zborowski, J. Xuetao [et al.]. URL: https://doi.org/10.1016/j.procs.2022.09.501 (2022 р.)
9. Telecommuting and food E-commerce: Socially sustainable practices during the COVID-19 pandemic in Canada / J. Music, S. Charlebois, V. Toole, C. Large. URL: https://doi.org/10.1016/j.trip.2021.100513 (2021 р.)
10. Has COVID-19 accelerated the E-commerce of agricultural products? Evidence from sales data of E-stores in China / J. Guo, S. Jin, J. Zhao [et al.]. URL: https://doi.org/10.1016/j.foodpol.2022.102377 (2022 р.)
11. Bilińska-Reformat K., Dewalska-Opitek A. Ecommerce as the predominant business model of fast fashion retailers in the era of global COVID 19 pandemics. URL: https://doi.org/10.1016/j.procs.2021.09.017 (2021 р.)
12. Kaplan M., Hotle S., Heaslip K. How has COVID-19 impacted customer perceptions and demand for delivery services: An exploratory analysis. URL: https://doi.org/10.1016/j.tranpol.2023.02.020 (2023 р.)
13. A literature review of the main factors influencing the e-commerce and last-mile delivery projects during COVID-19 pandemic / T. Campisi, A. Russo, S. Basbas [et al]. URL: https://doi.org/10.1016/j.trpro.2023.02.207 (2023 р.)
14. What matters for consumer sentiment in the euro area? World crude oil price or retail gasoline price? / S. Clerides, S. I. Krokida, N. Lambertides, D. Tsouknidis. URL: https://doi.org/10.1016/j.eneco.2021.105743 (2021 р.)
15. Ciani A., Mau K. Delivery times in international competition: An empirical investigation. URL: https://doi.org/10.1016/j.jdeveco.2022.103017 (2022 р.)
16. Hu J., Haddud A. Exploring the Impact of Globalization and Technology on Supply Chain Management: A Case of International E-Commerce Business.
URL: https://www.researchgate.net/publica -tion/338308216_Exploring_the_Impact_of_Globalization_and_Technology_on_Supply_Chain_Management_A_Case_of_International_E-Commerce_Business.