Data Mining in Sustainable Logistics

A survey on big data in best-practice companies [1] shows that only 14 percent of surveyed companies had a specific big data strategy, while 23 percent were planning to develop one in the future and 63 percent have no strategy. The quote ''Computers are useless. They can only give you answers." by Picasso, points out a big challenge in the adoption of big data. As data technologies such as data mining advance, it is still a challenge for other sectors to find out how they can benefit from this, and what the right question to ask is. The logistics sector is no exception. In ancient Greek roots, logistics means practical arithmetic which also indicates that data is indispensable in logistics management [2]. The flow of goods creates massive data sets to track shipments across the global network: origin and destination, size, weight, content, and location are all tracked [2]. Negative effects of city distribution such as air pollution, noise pollution and road safety are exerting pressure on cities. The question is if and how data mining can help to create a more sustainable approach to city logistics and at the same time increase the efficiency of the last mile. In this research, in order to provide a framework for the logistics sector to benefit from data mining, a thorough literature review on the application of data mining in sustainable logistics has been done. The concepts are classified into 3 groups introduced by Macharis et al. in [3]: Awareness, Avoidance and Act and Shift. Awareness emphasizes on tools quantifying the external effects and costs of city distribution in order to create awareness, which will lead to more sustainable solutions [3]. Avoidance focuses on possible consolidation solutions such as collaboration to avoid unnecessary vehicle movement and bundling goods together so that transport kilometres can be avoided [3]. Act and shift examines shifting to more environmentally friendly transport modes (barge, rail, tram, and cargo bikes) and electric vehicles and trucks [3]. It is important to raise awareness regarding the effect of city distribution. Data mining techniques can assist in performance analyses of logistics operations. Forecast models can be set up as a base reference for organisations. Furthermore, the role of data mining in external cost calculators and social cost-benefit analyses is explored. The state of the art of data mining applications in these fields will be further discussed. Next, unnecessary vehicle kilometres should be avoided. The vehicle-routing problem, warehouse management, location planning, batching, quality control and demand forecast are concepts which can greatly benefit from data mining techniques. Improvement in these concepts lead to reduction of travelled kilometres. A shift to more environmentally friendly modes of transport can diminish the negative impact of the kilometres which have to be driven. How data mining can improve transport mode shift to bike, truck platoons and barge will be discussed. Moreover, electric vehicles and self-driving cars apply a variety of data mining techniques, which will be analysed in details.References [1] C. Bange, T. Grosser, and N. Janoschek. Big data survey Europe: Usage, technology and budgets in European best practice companies. Würzburg: BARC Institute, 2013. [2] M. Jeseke, M. Grüner, and F. Wieß. Big data in logistics: A DHL perspective on how to move beyond the hype. DHL Customer Solutions & Innovation, December, 2013. [3] C. Macharis, S. Melo, J. Woxenius, and T. van Lier. Sustainable Logistics, volume 6. Emerald Group Publishing, 2014.