Transport is making its revolution at a time when the deliveries of goods and the transport of people are diversifying. How can AI responds to new challenges and increasingly complex issues? Alexandre Blouin, Head of Optimization and Operational Research at Maplink, answers in a forum.
By Alexandre BLOUIN
With the development of e-commerce and urban issues of sustainable development, the world of transport is evolving to adapt to new modes of delivery more numerous and more flexible, with an unprecedented volume of data to manage. Electric vehicles, autonomous vehicles, drones, … means of transport are in full revolution thanks to technological innovations that have reached the stage of maturity necessary to enter everyday life, both private and professional.
A few weeks ago, DHL Supply Chain announced the acquisition of several Tesla electric trucks, the manufacturer is also developing autonomous models. In the same period, Volvo and Uber presented their alliance to set up a fleet of autonomous taxis. In France, the RATP is currently testing the autonomous shuttle of the Toulouse startup Easymile. Another French startup, Navya, is also actively developing its taxi driverless models. What has been fiction for a long time is coming at a very high speed, and we will certainly see fleets of autonomous vehicles moving in all directions, from logistics to passenger transport.
Review of some strategic aspects:
1- The positioning of resources
The traditional taxi stations, where many cars wisely wait for customers to come to them – while these same customers often find it difficult to find one where they are, when they need it – is a model that should belong to the past.
The AI will look at forecasting to help decision-making in order to optimize the positioning of resources. These data will help to anticipate the places where there is the most likely demand, depending on the day of the week and the hours of the day, but also exogenous data such as weather, sports and cultural events, public transit strikes, etc.
This repositioning principle can be applied to all transport resources that are usually positioned on stations, such as Velib and Autolib. It can also apply to new self-service means of transport without a station, such as Cityscoot or Gobee.bike.
2- The assurance of being delivered on time
One of the key of performance and efficiency in transportation management is flexibility. Flows and demand are constantly changing, and the calculation of forecasts is hindered by the complexity and the data volumes to be taken into account.
It is important to remember that although a delivery is often free to the consumer, it is a cost to the business. To minimize costs, it is necessary to anticipate delivery peaks, for example during sales, to better size and allocate resources to absorb deliveries. The on-demand economy involves more and more frequent deliveries and multiple delivery options. To limit the costs, the company can orient the consumer on the most economically advantageous option compared to the available resources. The AI makes it possible to anticipate according to the deadlines and the points of delivery (relay points, home delivery, concierge, …).
One of the great strengths of the AI is its ability to handle huge volumes of data and adapt continuously, thanks to machine learning. It can therefore help to size fleets and fleets, including leased fleets. It can also promote better scheduling of drivers’ working hours.
3- The orchestration of deliveries
The management of delivery routes is a complex task that is often the subject of significant optimization margins but little exploited. And even more so when it comes to individual deliveries on multiple points, and that the time required by the consumers always shorten more (e-commerce deliveries, typically, or the deliveries of meals at home which meet a very successful).
It only takes a simple example to realize the importance of these optimization algorithms in the current context. Let’s take a fast food chain that receives hundreds of orders a day, and is committed to delivering all of its customers within 40 minutes. The AI and the machine learning make it possible to anticipate the peaks of requests, but also to predict which categories of products will be more or less ordered according to periods and exogenous factors. The AI may for example advise to wait before launching a route because it will anticipate other imminent orders, thus optimizing both the filling of trucks and routes.
These few examples show that the scope of the IA is particularly wide in terms of transport management. Assuredly, the AI will bring very significant benefits in the optimization of the transport, in particular thanks to its capacity to exploit large volumes of data for even more reliable forecasts.
Parallel to this revolution, flows of goods are exploding, driven in particular by the uninterrupted growth of e-commerce. The EFT Supply Chain Report 2017 highlights the challenges and development axes of professionals in this sector. The priorities are for 20,7% of them the costs of deliveries and, for 17%, the options of delivery. These issues are directly related to the ecological challenges, and the generalization of logistics on demand, we must find new delivery options especially to perform the “last mile”. The main development axes of the future Supply Chain are the cost reduction, the variety, the flexibility and the reliability of the new services.
Faced with this dual situation of transport means and intensification of market pressure, industry professionals must find answers in terms of management, management and optimization.
As in so many areas, artificial intelligence will address the challenge of complexity by providing a decision support tool that provides reliable indicators to improve logistics, leveraging large volumes of data to optimize delivery.